[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6MF456WWR41DHSFK25GDEPQ":3,"skill-same:01M6MF456WWR41DHSFK25GDEPQ":74},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":22,"publishedAt":43,"linkOnly":44,"markdownContent":45,"markdownBody":46,"frontMatter":47,"refName":54,"commitSha":55,"skillDirectory":56,"forksCount":13,"repositoryHtmlUrl":57,"repositoryHomepageUrl":6,"ownerHtmlUrl":58,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":59,"install":60,"resources":73},"01M6MF456WWR41DHSFK25GDEPQ","agent-protocol",null,"Inter-agent communication protocol for C-suite agent teams. Defines invocation syntax, loop prevention, isolation rules, and response formats. Use when C-suite agents need to query each other, coordinate cross-functional analysis, or run board meetings with multiple agent roles.","JiayiV\u002Fquestion_mixing","JiayiV","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F54466889?v=4","01M6JFJYYMRZ3QRGP9339JNQ2K","01M6JFJZ35KB0QVT50NAHPJA7G",0,"MIT","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Fpublic-02-claude-skills-main-gemini-skills-agent-protocol-skill-849686e1\u002FSKILL.md",[17],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42],"protocol","agent protocol","coordinate","suite","functional analysis","formats","agent-orchestration","agent communication protocol","functional","isolation","query","patterns","orchestration","agent teams","syntax","loop","analysis","prevention","communication","run","2026-11-10 10:12:30",false,"---\nname: \"agent-protocol\"\ndescription: \"Inter-agent communication protocol for C-suite agent teams. Defines invocation syntax, loop prevention, isolation rules, and response formats. Use when C-suite agents need to query each other, coordinate cross-functional analysis, or run board meetings with multiple agent roles.\"\nlicense: MIT\nmetadata:\n  version: 1.0.0\n  author: Alireza Rezvani\n  category: c-level\n  domain: agent-orchestration\n  updated: 2026-03-05\n  frameworks: invocation-patterns\n---\n\n# Inter-Agent Protocol\n\nHow C-suite agents talk to each other. Rules that prevent chaos, loops, and circular reasoning.\n\n## Keywords\nagent protocol, inter-agent communication, agent invocation, agent orchestration, multi-agent, c-suite coordination, agent chain, loop prevention, agent isolation, board meeting protocol\n\n## Invocation Syntax\n\nAny agent can query another using:\n\n```\n[INVOKE:role|question]\n```\n\n**Examples:**\n```\n[INVOKE:cfo|What's the burn rate impact of hiring 5 engineers in Q3?]\n[INVOKE:cto|Can we realistically ship this feature by end of quarter?]\n[INVOKE:chro|What's our typical time-to-hire for senior engineers?]\n[INVOKE:cro|What does our pipeline look like for the next 90 days?]\n```\n\n**Valid roles:** `ceo`, `cfo`, `cro`, `cmo`, `cpo`, `cto`, `chro`, `coo`, `ciso`, `gc`, `cdo`, `caio`, `cco`, `vpe`\n\n| Role token | Advisor skill |\n|---|---|\n| `gc` | general-counsel-advisor (legal, contracts, term sheets) |\n| `cdo` | chief-data-officer-advisor (data strategy, training-data rights) |\n| `caio` | chief-ai-officer-advisor (AI strategy, evals, AI risk) |\n| `cco` | chief-customer-officer-advisor (retention, customer success) |\n| `vpe` | vpe-advisor (engineering delivery, DORA, eng hiring) |\n\n## Response Format\n\nInvoked agents respond using this structure:\n\n```\n[RESPONSE:role]\nKey finding: [one line — the actual answer]\nSupporting data:\n  - [data point 1]\n  - [data point 2]\n  - [data point 3 — optional]\nConfidence: [high | medium | low]\nCaveat: [one line — what could make this wrong]\n[\u002FRESPONSE]\n```\n\n**Example:**\n```\n[RESPONSE:cfo]\nKey finding: Hiring 5 engineers in Q3 extends runway from 14 to 9 months at current burn.\nSupporting data:\n  - Current monthly burn: $280K → increases to ~$380K (+$100K fully loaded)\n  - ARR needed to offset: ~$1.2M additional within 12 months\n  - Current pipeline covers 60% of that target\nConfidence: medium\nCaveat: Assumes 3-month ramp and no change in revenue trajectory.\n[\u002FRESPONSE]\n```\n\n## Loop Prevention (Hard Rules)\n\nThese rules are enforced unconditionally. No exceptions.\n\n### Rule 1: No Self-Invocation\nAn agent cannot invoke itself.\n```\n❌ CFO → [INVOKE:cfo|...] — BLOCKED\n```\n\n### Rule 2: Maximum Depth = 2\nChains can go A→B→C. The third hop is blocked.\n```\n✅ CRO → CFO → COO (depth 2)\n❌ CRO → CFO → COO → CHRO (depth 3 — BLOCKED)\n```\n\n### Rule 3: No Circular Calls\nIf agent A called agent B, agent B cannot call agent A in the same chain.\n```\n✅ CRO → CFO → CMO\n❌ CRO → CFO → CRO (circular — BLOCKED)\n```\n\n### Rule 4: Chain Tracking\nEach invocation carries its call chain. Format:\n```\n[CHAIN: cro → cfo → coo]\n```\nAgents check this chain before responding with another invocation.\n\n**When blocked:** Return this instead of invoking:\n```\n[BLOCKED: cannot invoke cfo — circular call detected in chain cro→cfo]\nState assumption used instead: [explicit assumption the agent is making]\n```\n\n## Isolation Rules\n\n### Board Meeting Phase 2 (Independent Analysis)\n**NO invocations allowed.** Each role forms independent views before cross-pollination.\n- Reason: prevent anchoring and groupthink\n- Duration: entire Phase 2 analysis period\n- If an agent needs data from another role: state explicit assumption, flag it with `[ASSUMPTION: ...]`\n\n### Board Meeting Phase 3 (Critic Role)\nExecutive Mentor can **reference** other roles' outputs but **cannot invoke** them.\n- Reason: critique must be independent of new data requests\n- Allowed: \"The CFO's projection assumes X, which contradicts the CRO's pipeline data\"\n- Not allowed: `[INVOKE:cfo|...]` during critique phase\n\n### Outside Board Meetings\nInvocations are allowed freely, subject to loop prevention rules above.\n\n## When to Invoke vs When to Assume\n\n**Invoke when:**\n- The question requires domain-specific data you don't have\n- An error here would materially change the recommendation\n- The question is cross-functional by nature (e.g., hiring impact on both budget and capacity)\n\n**Assume when:**\n- The data is directionally clear and precision isn't critical\n- You're in Phase 2 isolation (always assume, never invoke)\n- The chain is already at depth 2\n- The question is minor compared to your main analysis\n\n**When assuming, always state it:**\n```\n[ASSUMPTION: runway ~12 months based on typical Series A burn profile — not verified with CFO]\n```\n\n## Conflict Resolution\n\nWhen two invoked agents give conflicting answers:\n\n1. **Flag the conflict explicitly:**\n   ```\n   [CONFLICT: CFO projects 14-month runway; CRO expects pipeline to close 80% → implies 18+ months]\n   ```\n2. **State the resolution approach:**\n   - Conservative: use the worse case\n   - Probabilistic: weight by confidence scores\n   - Escalate: flag for human decision\n3. **Never silently pick one** — surface the conflict to the user.\n\n## Broadcast Pattern (Crisis \u002F CEO)\n\nCEO can broadcast to all roles simultaneously:\n```\n[BROADCAST:all|What's the impact if we miss the fundraise?]\n```\n\nResponses come back independently (no agent sees another's response before forming its own). Aggregate after all respond.\n\n## Decision Memory (Canonical Layout)\n\nAll C-suite skills and `\u002Fcs:*` commands read and write decisions in **one** place — the two-layer model owned by `\u002Fcs:decide` and the decision-logger skill:\n\n```\n~\u002F.claude\u002Fdecisions\u002F\n├── raw\u002FYYYY-MM-DD-\u003Cslug>.md        # Layer 1 — full transcripts\u002Fdeliberations (never auto-loaded)\n├── raw\u002Farchive\u002FYYYY\u002F               # Raw files after 90 days\n├── approved\u002FYYYY-MM-DD-\u003Cslug>.md   # Layer 2 — one founder-approved decision record per file\n└── approved\u002Fdecisions.md           # Layer 2 index — append-only log of approved decisions\n```\n\n**Rules:**\n- **Layer 1 (raw)** stores everything, including rejected arguments. Reference only — never feeds future sessions automatically.\n- **Layer 2 (approved)** stores only founder-approved decisions. This is what board meetings, `\u002Fcs:office-hours`, and `\u002Fcs:founder-mode` load. Prevents hallucinated consensus.\n- Writers: `\u002Fcs:decide` and the Chief of Staff (post board-meeting Phase 5). Individual role agents never write decisions directly.\n- decision-logger, chief-of-staff, and board-meeting all use this layout. Their SKILL.md files link here rather than defining their own paths.\n\n**Migration:** earlier versions used `memory\u002Fboard-meetings\u002F` (decision-logger, board-meeting) and `~\u002F.claude\u002Fdecision-log.md` (chief-of-staff); read those for history if present, but write all new entries to `~\u002F.claude\u002Fdecisions\u002F`.\n\n## Quick Reference\n\n| Rule | Behavior |\n|------|----------|\n| Self-invoke | ❌ Always blocked |\n| Depth > 2 | ❌ Blocked, state assumption |\n| Circular | ❌ Blocked, state assumption |\n| Phase 2 isolation | ❌ No invocations |\n| Phase 3 critique | ❌ Reference only, no invoke |\n| Conflict | ✅ Surface it, don't hide it |\n| Assumption | ✅ Always explicit with `[ASSUMPTION: ...]` |\n\n## Internal Quality Loop (before anything reaches the founder)\n\nNo role presents to the founder without passing through this verification loop. The founder sees polished, verified output — not first drafts.\n\n### Step 1: Self-Verification (every role, every time)\n\nBefore presenting, every role runs this internal checklist:\n\n```\nSELF-VERIFY CHECKLIST:\n□ Source Attribution — Where did each data point come from?\n  ✅ \"ARR is $2.1M (from CRO pipeline report, Q4 actuals)\"\n  ❌ \"ARR is around $2M\" (no source, vague)\n\n□ Assumption Audit — What am I assuming vs what I verified?\n  Tag every assumption: [VERIFIED: checked against data] or [ASSUMED: not verified]\n  If >50% of findings are ASSUMED → flag low confidence\n\n□ Confidence Score — How sure am I on each finding?\n  ? High: verified data, established pattern, multiple sources\n  ? Medium: single source, reasonable inference, some uncertainty\n  ? Low: assumption-based, limited data, first-time analysis\n\n□ Contradiction Check — Does this conflict with known context?\n  Check against company-context.md and recent decisions in decision-log\n  If it contradicts a past decision → flag explicitly\n\n□ \"So What?\" Test — Does every finding have a business consequence?\n  If you can't answer \"so what?\" in one sentence → cut it\n```\n\n### Step 2: Peer Verification (cross-functional validation)\n\nWhen a recommendation impacts another role's domain, that role validates BEFORE presenting.\n\n| If your recommendation involves... | Validate with... | They check... |\n|-------------------------------------|-------------------|---------------|\n| Financial numbers or budget | CFO | Math, runway impact, budget reality |\n| Revenue projections | CRO | Pipeline backing, historical accuracy |\n| Headcount or hiring | CHRO | Market reality, comp feasibility, timeline |\n| Technical feasibility or timeline | CTO | Engineering capacity, technical debt load |\n| Operational process changes | COO | Capacity, dependencies, scaling impact |\n| Customer-facing changes | CRO + CPO | Churn risk, product roadmap conflict |\n| Security or compliance claims | CISO | Actual posture, regulation requirements |\n| Market or positioning claims | CMO | Data backing, competitive reality |\n| Legal exposure, contracts, term sheets | GC | Clause risk, IP ownership, regulatory triggers |\n| Data rights, training-data provenance | CDO | Consent basis, GDPR Art. 6, data-asset impact |\n| AI model claims, eval results, AI risk | CAIO | Eval coverage, hallucination SLO, EU AI Act tier |\n| Retention, churn, customer-health claims | CCO | GRR\u002FNRR decomposition, churn root cause |\n| Delivery timelines, eng throughput | VPE | DORA metrics, cycle-time reality, team capacity |\n\n**Peer validation format:**\n```\n[PEER-VERIFY:cfo]\nValidated: ✅ Burn rate calculation correct\nAdjusted: ⚠️ Hiring timeline should be Q3 not Q2 (budget constraint)\nFlagged: ? Missing equity cost in total comp projection\n[\u002FPEER-VERIFY]\n```\n\n**Skip peer verification when:**\n- Single-domain question with no cross-functional impact\n- Time-sensitive proactive alert (send alert, verify after)\n- Founder explicitly asked for a quick take\n\n### Step 3: Critic Pre-Screen (high-stakes decisions only)\n\nFor decisions that are **irreversible, high-cost, or bet-the-company**, the Executive Mentor pre-screens before the founder sees it.\n\n**Triggers for pre-screen:**\n- Involves spending > 20% of remaining runway\n- Affects >30% of the team (layoffs, reorg)\n- Changes company strategy or direction\n- Involves external commitments (fundraising terms, partnerships, M&A)\n- Any recommendation where all roles agree (suspicious consensus)\n\n**Pre-screen output:**\n```\n[CRITIC-SCREEN]\nWeakest point: [The single biggest vulnerability in this recommendation]\nMissing perspective: [What nobody considered]\nIf wrong, the cost is: [Quantified downside]\nProceed: ✅ With noted risks | ⚠️ After addressing [specific gap] | ? Rethink\n[\u002FCRITIC-SCREEN]\n```\n\n### Step 4: Course Correction (after founder feedback)\n\nThe loop doesn't end at delivery. After the founder responds:\n\n```\nFOUNDER FEEDBACK LOOP:\n1. Founder approves → log decision (Layer 2), assign actions\n2. Founder modifies → update analysis with corrections, re-verify changed parts\n3. Founder rejects → log rejection with DO_NOT_RESURFACE, understand WHY\n4. Founder asks follow-up → deepen analysis on specific point, re-verify\n\nPOST-DECISION REVIEW (30\u002F60\u002F90 days):\n- Was the recommendation correct?\n- What did we miss?\n- Update company-context.md with what we learned\n- If wrong → document the lesson, adjust future analysis\n```\n\n### Verification Level by Stakes\n\n| Stakes | Self-Verify | Peer-Verify | Critic Pre-Screen |\n|--------|-------------|-------------|-------------------|\n| Low (informational) | ✅ Required | ❌ Skip | ❌ Skip |\n| Medium (operational) | ✅ Required | ✅ Required | ❌ Skip |\n| High (strategic) | ✅ Required | ✅ Required | ✅ Required |\n| Critical (irreversible) | ✅ Required | ✅ Required | ✅ Required + board meeting |\n\n### What Changes in the Output Format\n\nThe verified output adds confidence and source information:\n\n```\nBOTTOM LINE\n[Answer] — Confidence: ? High\n\nWHAT\n• [Finding 1] [VERIFIED: Q4 actuals] ?\n• [Finding 2] [VERIFIED: CRO pipeline data] ?  \n• [Finding 3] [ASSUMED: based on industry benchmarks] ?\n\nPEER-VERIFIED BY: CFO (math ✅), CTO (timeline ⚠️ adjusted to Q3)\n```\n\n---\n\n## User Communication Standard\n\nAll C-suite output to the founder follows ONE format. No exceptions. The founder is the decision-maker — give them results, not process.\n\n### Standard Output (single-role response)\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? [ROLE] — [Topic]\n\nBOTTOM LINE\n[One sentence. The answer. No preamble.]\n\nWHAT\n• [Finding 1 — most critical]\n• [Finding 2]\n• [Finding 3]\n(Max 5 bullets. If more needed → reference doc.)\n\nWHY THIS MATTERS\n[1-2 sentences. Business impact. Not theory — consequence.]\n\nHOW TO ACT\n1. [Action] → [Owner] → [Deadline]\n2. [Action] → [Owner] → [Deadline]\n3. [Action] → [Owner] → [Deadline]\n\n⚠️ RISKS (if any)\n• [Risk + what triggers it]\n\n? YOUR DECISION (if needed)\nOption A: [Description] — [Trade-off]\nOption B: [Description] — [Trade-off]\nRecommendation: [Which and why, in one line]\n\n? DETAIL: [reference doc or script output for deep-dive]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n### Proactive Alert (unsolicited — triggered by context)\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? [ROLE] — Proactive Alert\n\nWHAT I NOTICED\n[What triggered this — specific, not vague]\n\nWHY IT MATTERS\n[Business consequence if ignored — in dollars, time, or risk]\n\nRECOMMENDED ACTION\n[Exactly what to do, who does it, by when]\n\nURGENCY: ? Act today | ? This week | ⚪ Next review\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n### Board Meeting Output (multi-role synthesis)\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? BOARD MEETING — [Date] — [Agenda Topic]\n\nDECISION REQUIRED\n[Frame the decision in one sentence]\n\nPERSPECTIVES\n  CEO: [one-line position]\n  CFO: [one-line position]\n  CRO: [one-line position]\n  [... only roles that contributed]\n\nWHERE THEY AGREE\n• [Consensus point 1]\n• [Consensus point 2]\n\nWHERE THEY DISAGREE\n• [Conflict] — CEO says X, CFO says Y\n• [Conflict] — CRO says X, CPO says Y\n\nCRITIC'S VIEW (Executive Mentor)\n[The uncomfortable truth nobody else said]\n\nRECOMMENDED DECISION\n[Clear recommendation with rationale]\n\nACTION ITEMS\n1. [Action] → [Owner] → [Deadline]\n2. [Action] → [Owner] → [Deadline]\n3. [Action] → [Owner] → [Deadline]\n\n? YOUR CALL\n[Options if you disagree with the recommendation]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n### Communication Rules (non-negotiable)\n\n1. **Bottom line first.** Always. The founder's time is the scarcest resource.\n2. **Results and decisions only.** No process narration (\"First I analyzed...\"). No thinking out loud.\n3. **What + Why + How.** Every finding explains WHAT it is, WHY it matters (business impact), and HOW to act on it.\n4. **Max 5 bullets per section.** Longer = reference doc.\n5. **Actions have owners and deadlines.** \"We should consider\" is banned. Who does what by when.\n6. **Decisions framed as options.** Not \"what do you think?\" — \"Option A or B, here's the trade-off, here's my recommendation.\"\n7. **The founder decides.** Roles recommend. The founder approves, modifies, or rejects. Every output respects this hierarchy.\n8. **Risks are concrete.** Not \"there might be risks\" — \"if X happens, Y breaks, costing $Z.\"\n9. **No jargon without explanation.** If you use a term, explain it on first use.\n10. **Silence is an option.** If there's nothing to report, don't fabricate updates.\n\n## Reference\n- `references\u002Finvocation-patterns.md` — common cross-functional patterns with examples\n","# Inter-Agent Protocol\n\nHow C-suite agents talk to each other. Rules that prevent chaos, loops, and circular reasoning.\n\n## Keywords\nagent protocol, inter-agent communication, agent invocation, agent orchestration, multi-agent, c-suite coordination, agent chain, loop prevention, agent isolation, board meeting protocol\n\n## Invocation Syntax\n\nAny agent can query another using:\n\n```\n[INVOKE:role|question]\n```\n\n**Examples:**\n```\n[INVOKE:cfo|What's the burn rate impact of hiring 5 engineers in Q3?]\n[INVOKE:cto|Can we realistically ship this feature by end of quarter?]\n[INVOKE:chro|What's our typical time-to-hire for senior engineers?]\n[INVOKE:cro|What does our pipeline look like for the next 90 days?]\n```\n\n**Valid roles:** `ceo`, `cfo`, `cro`, `cmo`, `cpo`, `cto`, `chro`, `coo`, `ciso`, `gc`, `cdo`, `caio`, `cco`, `vpe`\n\n| Role token | Advisor skill |\n|---|---|\n| `gc` | general-counsel-advisor (legal, contracts, term sheets) |\n| `cdo` | chief-data-officer-advisor (data strategy, training-data rights) |\n| `caio` | chief-ai-officer-advisor (AI strategy, evals, AI risk) |\n| `cco` | chief-customer-officer-advisor (retention, customer success) |\n| `vpe` | vpe-advisor (engineering delivery, DORA, eng hiring) |\n\n## Response Format\n\nInvoked agents respond using this structure:\n\n```\n[RESPONSE:role]\nKey finding: [one line — the actual answer]\nSupporting data:\n  - [data point 1]\n  - [data point 2]\n  - [data point 3 — optional]\nConfidence: [high | medium | low]\nCaveat: [one line — what could make this wrong]\n[\u002FRESPONSE]\n```\n\n**Example:**\n```\n[RESPONSE:cfo]\nKey finding: Hiring 5 engineers in Q3 extends runway from 14 to 9 months at current burn.\nSupporting data:\n  - Current monthly burn: $280K → increases to ~$380K (+$100K fully loaded)\n  - ARR needed to offset: ~$1.2M additional within 12 months\n  - Current pipeline covers 60% of that target\nConfidence: medium\nCaveat: Assumes 3-month ramp and no change in revenue trajectory.\n[\u002FRESPONSE]\n```\n\n## Loop Prevention (Hard Rules)\n\nThese rules are enforced unconditionally. No exceptions.\n\n### Rule 1: No Self-Invocation\nAn agent cannot invoke itself.\n```\n❌ CFO → [INVOKE:cfo|...] — BLOCKED\n```\n\n### Rule 2: Maximum Depth = 2\nChains can go A→B→C. The third hop is blocked.\n```\n✅ CRO → CFO → COO (depth 2)\n❌ CRO → CFO → COO → CHRO (depth 3 — BLOCKED)\n```\n\n### Rule 3: No Circular Calls\nIf agent A called agent B, agent B cannot call agent A in the same chain.\n```\n✅ CRO → CFO → CMO\n❌ CRO → CFO → CRO (circular — BLOCKED)\n```\n\n### Rule 4: Chain Tracking\nEach invocation carries its call chain. Format:\n```\n[CHAIN: cro → cfo → coo]\n```\nAgents check this chain before responding with another invocation.\n\n**When blocked:** Return this instead of invoking:\n```\n[BLOCKED: cannot invoke cfo — circular call detected in chain cro→cfo]\nState assumption used instead: [explicit assumption the agent is making]\n```\n\n## Isolation Rules\n\n### Board Meeting Phase 2 (Independent Analysis)\n**NO invocations allowed.** Each role forms independent views before cross-pollination.\n- Reason: prevent anchoring and groupthink\n- Duration: entire Phase 2 analysis period\n- If an agent needs data from another role: state explicit assumption, flag it with `[ASSUMPTION: ...]`\n\n### Board Meeting Phase 3 (Critic Role)\nExecutive Mentor can **reference** other roles' outputs but **cannot invoke** them.\n- Reason: critique must be independent of new data requests\n- Allowed: \"The CFO's projection assumes X, which contradicts the CRO's pipeline data\"\n- Not allowed: `[INVOKE:cfo|...]` during critique phase\n\n### Outside Board Meetings\nInvocations are allowed freely, subject to loop prevention rules above.\n\n## When to Invoke vs When to Assume\n\n**Invoke when:**\n- The question requires domain-specific data you don't have\n- An error here would materially change the recommendation\n- The question is cross-functional by nature (e.g., hiring impact on both budget and capacity)\n\n**Assume when:**\n- The data is directionally clear and precision isn't critical\n- You're in Phase 2 isolation (always assume, never invoke)\n- The chain is already at depth 2\n- The question is minor compared to your main analysis\n\n**When assuming, always state it:**\n```\n[ASSUMPTION: runway ~12 months based on typical Series A burn profile — not verified with CFO]\n```\n\n## Conflict Resolution\n\nWhen two invoked agents give conflicting answers:\n\n1. **Flag the conflict explicitly:**\n   ```\n   [CONFLICT: CFO projects 14-month runway; CRO expects pipeline to close 80% → implies 18+ months]\n   ```\n2. **State the resolution approach:**\n   - Conservative: use the worse case\n   - Probabilistic: weight by confidence scores\n   - Escalate: flag for human decision\n3. **Never silently pick one** — surface the conflict to the user.\n\n## Broadcast Pattern (Crisis \u002F CEO)\n\nCEO can broadcast to all roles simultaneously:\n```\n[BROADCAST:all|What's the impact if we miss the fundraise?]\n```\n\nResponses come back independently (no agent sees another's response before forming its own). Aggregate after all respond.\n\n## Decision Memory (Canonical Layout)\n\nAll C-suite skills and `\u002Fcs:*` commands read and write decisions in **one** place — the two-layer model owned by `\u002Fcs:decide` and the decision-logger skill:\n\n```\n~\u002F.claude\u002Fdecisions\u002F\n├── raw\u002FYYYY-MM-DD-\u003Cslug>.md        # Layer 1 — full transcripts\u002Fdeliberations (never auto-loaded)\n├── raw\u002Farchive\u002FYYYY\u002F               # Raw files after 90 days\n├── approved\u002FYYYY-MM-DD-\u003Cslug>.md   # Layer 2 — one founder-approved decision record per file\n└── approved\u002Fdecisions.md           # Layer 2 index — append-only log of approved decisions\n```\n\n**Rules:**\n- **Layer 1 (raw)** stores everything, including rejected arguments. Reference only — never feeds future sessions automatically.\n- **Layer 2 (approved)** stores only founder-approved decisions. This is what board meetings, `\u002Fcs:office-hours`, and `\u002Fcs:founder-mode` load. Prevents hallucinated consensus.\n- Writers: `\u002Fcs:decide` and the Chief of Staff (post board-meeting Phase 5). Individual role agents never write decisions directly.\n- decision-logger, chief-of-staff, and board-meeting all use this layout. Their SKILL.md files link here rather than defining their own paths.\n\n**Migration:** earlier versions used `memory\u002Fboard-meetings\u002F` (decision-logger, board-meeting) and `~\u002F.claude\u002Fdecision-log.md` (chief-of-staff); read those for history if present, but write all new entries to `~\u002F.claude\u002Fdecisions\u002F`.\n\n## Quick Reference\n\n| Rule | Behavior |\n|------|----------|\n| Self-invoke | ❌ Always blocked |\n| Depth > 2 | ❌ Blocked, state assumption |\n| Circular | ❌ Blocked, state assumption |\n| Phase 2 isolation | ❌ No invocations |\n| Phase 3 critique | ❌ Reference only, no invoke |\n| Conflict | ✅ Surface it, don't hide it |\n| Assumption | ✅ Always explicit with `[ASSUMPTION: ...]` |\n\n## Internal Quality Loop (before anything reaches the founder)\n\nNo role presents to the founder without passing through this verification loop. The founder sees polished, verified output — not first drafts.\n\n### Step 1: Self-Verification (every role, every time)\n\nBefore presenting, every role runs this internal checklist:\n\n```\nSELF-VERIFY CHECKLIST:\n□ Source Attribution — Where did each data point come from?\n  ✅ \"ARR is $2.1M (from CRO pipeline report, Q4 actuals)\"\n  ❌ \"ARR is around $2M\" (no source, vague)\n\n□ Assumption Audit — What am I assuming vs what I verified?\n  Tag every assumption: [VERIFIED: checked against data] or [ASSUMED: not verified]\n  If >50% of findings are ASSUMED → flag low confidence\n\n□ Confidence Score — How sure am I on each finding?\n  ? High: verified data, established pattern, multiple sources\n  ? Medium: single source, reasonable inference, some uncertainty\n  ? Low: assumption-based, limited data, first-time analysis\n\n□ Contradiction Check — Does this conflict with known context?\n  Check against company-context.md and recent decisions in decision-log\n  If it contradicts a past decision → flag explicitly\n\n□ \"So What?\" Test — Does every finding have a business consequence?\n  If you can't answer \"so what?\" in one sentence → cut it\n```\n\n### Step 2: Peer Verification (cross-functional validation)\n\nWhen a recommendation impacts another role's domain, that role validates BEFORE presenting.\n\n| If your recommendation involves... | Validate with... | They check... |\n|-------------------------------------|-------------------|---------------|\n| Financial numbers or budget | CFO | Math, runway impact, budget reality |\n| Revenue projections | CRO | Pipeline backing, historical accuracy |\n| Headcount or hiring | CHRO | Market reality, comp feasibility, timeline |\n| Technical feasibility or timeline | CTO | Engineering capacity, technical debt load |\n| Operational process changes | COO | Capacity, dependencies, scaling impact |\n| Customer-facing changes | CRO + CPO | Churn risk, product roadmap conflict |\n| Security or compliance claims | CISO | Actual posture, regulation requirements |\n| Market or positioning claims | CMO | Data backing, competitive reality |\n| Legal exposure, contracts, term sheets | GC | Clause risk, IP ownership, regulatory triggers |\n| Data rights, training-data provenance | CDO | Consent basis, GDPR Art. 6, data-asset impact |\n| AI model claims, eval results, AI risk | CAIO | Eval coverage, hallucination SLO, EU AI Act tier |\n| Retention, churn, customer-health claims | CCO | GRR\u002FNRR decomposition, churn root cause |\n| Delivery timelines, eng throughput | VPE | DORA metrics, cycle-time reality, team capacity |\n\n**Peer validation format:**\n```\n[PEER-VERIFY:cfo]\nValidated: ✅ Burn rate calculation correct\nAdjusted: ⚠️ Hiring timeline should be Q3 not Q2 (budget constraint)\nFlagged: ? Missing equity cost in total comp projection\n[\u002FPEER-VERIFY]\n```\n\n**Skip peer verification when:**\n- Single-domain question with no cross-functional impact\n- Time-sensitive proactive alert (send alert, verify after)\n- Founder explicitly asked for a quick take\n\n### Step 3: Critic Pre-Screen (high-stakes decisions only)\n\nFor decisions that are **irreversible, high-cost, or bet-the-company**, the Executive Mentor pre-screens before the founder sees it.\n\n**Triggers for pre-screen:**\n- Involves spending > 20% of remaining runway\n- Affects >30% of the team (layoffs, reorg)\n- Changes company strategy or direction\n- Involves external commitments (fundraising terms, partnerships, M&A)\n- Any recommendation where all roles agree (suspicious consensus)\n\n**Pre-screen output:**\n```\n[CRITIC-SCREEN]\nWeakest point: [The single biggest vulnerability in this recommendation]\nMissing perspective: [What nobody considered]\nIf wrong, the cost is: [Quantified downside]\nProceed: ✅ With noted risks | ⚠️ After addressing [specific gap] | ? Rethink\n[\u002FCRITIC-SCREEN]\n```\n\n### Step 4: Course Correction (after founder feedback)\n\nThe loop doesn't end at delivery. After the founder responds:\n\n```\nFOUNDER FEEDBACK LOOP:\n1. Founder approves → log decision (Layer 2), assign actions\n2. Founder modifies → update analysis with corrections, re-verify changed parts\n3. Founder rejects → log rejection with DO_NOT_RESURFACE, understand WHY\n4. Founder asks follow-up → deepen analysis on specific point, re-verify\n\nPOST-DECISION REVIEW (30\u002F60\u002F90 days):\n- Was the recommendation correct?\n- What did we miss?\n- Update company-context.md with what we learned\n- If wrong → document the lesson, adjust future analysis\n```\n\n### Verification Level by Stakes\n\n| Stakes | Self-Verify | Peer-Verify | Critic Pre-Screen |\n|--------|-------------|-------------|-------------------|\n| Low (informational) | ✅ Required | ❌ Skip | ❌ Skip |\n| Medium (operational) | ✅ Required | ✅ Required | ❌ Skip |\n| High (strategic) | ✅ Required | ✅ Required | ✅ Required |\n| Critical (irreversible) | ✅ Required | ✅ Required | ✅ Required + board meeting |\n\n### What Changes in the Output Format\n\nThe verified output adds confidence and source information:\n\n```\nBOTTOM LINE\n[Answer] — Confidence: ? High\n\nWHAT\n• [Finding 1] [VERIFIED: Q4 actuals] ?\n• [Finding 2] [VERIFIED: CRO pipeline data] ?  \n• [Finding 3] [ASSUMED: based on industry benchmarks] ?\n\nPEER-VERIFIED BY: CFO (math ✅), CTO (timeline ⚠️ adjusted to Q3)\n```\n\n---\n\n## User Communication Standard\n\nAll C-suite output to the founder follows ONE format. No exceptions. The founder is the decision-maker — give them results, not process.\n\n### Standard Output (single-role response)\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? [ROLE] — [Topic]\n\nBOTTOM LINE\n[One sentence. The answer. No preamble.]\n\nWHAT\n• [Finding 1 — most critical]\n• [Finding 2]\n• [Finding 3]\n(Max 5 bullets. If more needed → reference doc.)\n\nWHY THIS MATTERS\n[1-2 sentences. Business impact. Not theory — consequence.]\n\nHOW TO ACT\n1. [Action] → [Owner] → [Deadline]\n2. [Action] → [Owner] → [Deadline]\n3. [Action] → [Owner] → [Deadline]\n\n⚠️ RISKS (if any)\n• [Risk + what triggers it]\n\n? YOUR DECISION (if needed)\nOption A: [Description] — [Trade-off]\nOption B: [Description] — [Trade-off]\nRecommendation: [Which and why, in one line]\n\n? DETAIL: [reference doc or script output for deep-dive]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n### Proactive Alert (unsolicited — triggered by context)\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? [ROLE] — Proactive Alert\n\nWHAT I NOTICED\n[What triggered this — specific, not vague]\n\nWHY IT MATTERS\n[Business consequence if ignored — in dollars, time, or risk]\n\nRECOMMENDED ACTION\n[Exactly what to do, who does it, by when]\n\nURGENCY: ? Act today | ? This week | ⚪ Next review\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n### Board Meeting Output (multi-role synthesis)\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? BOARD MEETING — [Date] — [Agenda Topic]\n\nDECISION REQUIRED\n[Frame the decision in one sentence]\n\nPERSPECTIVES\n  CEO: [one-line position]\n  CFO: [one-line position]\n  CRO: [one-line position]\n  [... only roles that contributed]\n\nWHERE THEY AGREE\n• [Consensus point 1]\n• [Consensus point 2]\n\nWHERE THEY DISAGREE\n• [Conflict] — CEO says X, CFO says Y\n• [Conflict] — CRO says X, CPO says Y\n\nCRITIC'S VIEW (Executive Mentor)\n[The uncomfortable truth nobody else said]\n\nRECOMMENDED DECISION\n[Clear recommendation with rationale]\n\nACTION ITEMS\n1. [Action] → [Owner] → [Deadline]\n2. [Action] → [Owner] → [Deadline]\n3. [Action] → [Owner] → [Deadline]\n\n? YOUR CALL\n[Options if you disagree with the recommendation]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n### Communication Rules (non-negotiable)\n\n1. **Bottom line first.** Always. The founder's time is the scarcest resource.\n2. **Results and decisions only.** No process narration (\"First I analyzed...\"). No thinking out loud.\n3. **What + Why + How.** Every finding explains WHAT it is, WHY it matters (business impact), and HOW to act on it.\n4. **Max 5 bullets per section.** Longer = reference doc.\n5. **Actions have owners and deadlines.** \"We should consider\" is banned. Who does what by when.\n6. **Decisions framed as options.** Not \"what do you think?\" — \"Option A or B, here's the trade-off, here's my recommendation.\"\n7. **The founder decides.** Roles recommend. The founder approves, modifies, or rejects. Every output respects this hierarchy.\n8. **Risks are concrete.** Not \"there might be risks\" — \"if X happens, Y breaks, costing $Z.\"\n9. **No jargon without explanation.** If you use a term, explain it on first use.\n10. **Silence is an option.** If there's nothing to report, don't fabricate updates.\n\n## Reference\n- `references\u002Finvocation-patterns.md` — common cross-functional patterns with examples\n",{"name":5,"license":14,"metadata":48,"description":7},{"author":49,"domain":29,"updated":50,"version":51,"category":52,"frameworks":53},"Alireza Rezvani",1772668800000,"1.0.0","c-level","invocation-patterns","main","068c2486ccc0f8d200a3db4d21ec137064eda3b4","skills\u002Fpublic-02-claude-skills-main-gemini-skills-agent-protocol-skill-849686e1","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing","https:\u002F\u002Fgithub.com\u002FJiayiV","2026-08-25 16:30:52",{"source":8,"skillName":5,"commands":61,"prompt":71,"zipAvailable":72},[62,65,68],{"tool":63,"command":64},"npx","npx skills add JiayiV\u002Fquestion_mixing --skill agent-protocol",{"tool":66,"command":67},"bunx","bunx skills add JiayiV\u002Fquestion_mixing --skill agent-protocol",{"tool":69,"command":70},"pnpm","pnpm dlx skills add JiayiV\u002Fquestion_mixing --skill agent-protocol","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Fpublic-02-claude-skills-main-gemini-skills-agent-protocol-skill-849686e1\u002FSKILL.md",true,[],[75,101,128,157,185,220],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":43},"01M6NN9PJS9NHWPDG4QP93BP8F","webmcpify","Make a web app agent-ready — propose a WebMCP tool manifest, integrate, verify in a real browser, heal; unrelated code stays untouched. Use for \"webmcpify\", \"add WebMCP\", or \"expose app actions to AI agents\".","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Fwebmcpify\u002FSKILL.md",[],[77,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100],"github","ai agents","web","notes","code","browser","https","actions","manifest","inventory","verify","tool","use","ready","unrelated","app","scope","for","webmcp",{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":105,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":43},"01M6NN6KZE5RBN6NGHN2WXV1RE","vardoger-analyze","Use when the user asks to personalize the GitHub Copilot CLI assistant, adapt Copilot to their style, use vardoger, or analyze their Copilot CLI conversation history. Reads the local session directory at `~\u002F.copilot\u002Fsession-state\u002F`, extracts recurring preferences and conventions, and writes a fenced personalization block into `~\u002F.copilot\u002Fcopilot-instructions.md`. Runs entirely on the user's machine via the local `vardoger` CLI (`pipx install vardoger`); no network calls and no uploads. Triggers: 'personalize my copilot', 'analyze my copilot history', 'tailor copilot to me', 'run vardoger', 'update my copilot instructions from history', 'make copilot learn my style'.","Apache-2.0","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Fvardoger-analyze\u002FSKILL.md",[],[109,110,111,112,113,114,42,115,116,117,118,119,120,121,122,123,124,125,126,127],"vardoger analyze","analyze","directory","cli","github copilot","copilot","update","learn","network","my style","install","conversation","me","local","github copilot cli","pipx","my","at","copilot instructions",{"skillNo":129,"name":130,"nameZh":6,"description":131,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":132,"categories":133,"keywords":138,"publishedAt":43},"01M6NN3CZGQP3XAJ04BM1NNYGT","transloadit-media-processing","Process media files (video, audio, images, documents) using Transloadit. Use when asked to encode video to HLS\u002FMP4, generate thumbnails, resize or watermark images, extract audio, concatenate clips, add subtitles, OCR documents, or run any media processing pipeline. Covers 86+ processing robots for file transformation at scale.","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Ftransloadit-media-processing\u002FSKILL.md",[134,137],{"categoryNo":135,"categoryName":136,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000016","图像与多媒体",{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},[139,140,141,142,143,144,145,146,147,148,149,150,151,94,152,153,154,155,99,156],"transloadit media processing","Media","media processing","processing","MCP","mcp server","documents","subtitles","pipeline","ocr","Video","Audio","mp4","images","files","transformation","free","hls",{"skillNo":158,"name":159,"nameZh":6,"description":160,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":161,"categories":162,"keywords":166,"publishedAt":43},"01M6NN28E92F9GH4WFYQ2H1Y15","terraform-azurerm-set-diff-analyzer","Analyze Terraform plan JSON output for AzureRM Provider to distinguish between false-positive diffs (order-only changes in Set-type attributes) and actual resource changes. Use when reviewing terraform plan output for Azure resources like Application Gateway, Load Balancer, Firewall, Front Door, NSG, and other resources with Set-type attributes that cause spurious diffs due to internal ordering changes.","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Fterraform-azurerm-set-diff-analyzer\u002FSKILL.md",[163],{"categoryNo":164,"categoryName":165,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000011","运维与部署",[167,168,169,170,171,172,173,174,175,94,176,177,178,179,180,181,182,183,184,99],"terraform","SET","terraform azurerm set diff analyzer","diff","analyzer","diffs","azure","gateway","firewall","only","resource changes","provider","resources","load balancer","attributes","front","load","balancer",{"skillNo":186,"name":187,"nameZh":6,"description":188,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":189,"categories":190,"keywords":202,"publishedAt":43},"01M6NN1QMH2TDYWTH3XY3N5XPV","technical-job-search","Use this skill when a software engineer asks for help with job search tasks: parsing or analyzing a job description, tailoring a CV\u002Fresume, writing a cover letter, evaluating a job offer, or drafting a post-interview follow-up email. Do not activate for general career advice unrelated to an active job search action.","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Ftechnical-job-search\u002FSKILL.md",[191,194,199],{"categoryNo":192,"categoryName":193,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":195,"categoryName":196,"taxonomyType":197,"taxonomyTypeDesc":198},"01JBSEEDCAT000000000000036","内容创作者","PROFESSION","适用职业",{"categoryNo":200,"categoryName":201,"taxonomyType":197,"taxonomyTypeDesc":198},"01JBSEEDCAT000000000000038","人力资源",[203,204,205,206,207,208,209,210,110,211,94,212,213,214,215,216,217,218,99,219],"search","Job Search","technical job search","technical","writing","tasks","email","parsing","interview","up","engineer","general","activate","do","task","cv","letter",{"skillNo":221,"name":222,"nameZh":6,"description":223,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":224,"categories":225,"keywords":226,"publishedAt":43},"01M6NMZFA67XYS1GS09WT42235","steno-mode","Shorthand-first response compression that cuts ~40% of response tokens while preserving technical precision and exact literals. Use when the user says \"steno mode\", \"shorthand mode\", \"compressed responses\", \"token reduction\", \"brief structured output\", or invokes \u002Fsteno. Supports four compression levels: lite, brief, court, machine. Do not trigger for requests needing polished prose such as onboarding\u002Ftutorial content, stakeholder or customer-facing copy, or teaching-focused explanations.","https:\u002F\u002Fgithub.com\u002FJiayiV\u002Fquestion_mixing\u002Fblob\u002Fmain\u002Fskills\u002Fsteno-mode\u002FSKILL.md",[],[227,228,229,230,231,206,232,94,216,233,234,235,236,237,238,239,240,241,99,242],"steno mode","mode","customer","brief","first","tutorial","teaching","tokens","as","trigger","structured output","machine","output","court","requests","response compression"]