qa-poc
Validates Proof of Concept completeness: success criteria coverage, evidence quality, TTV measurement, commercial validation, and Go/No-Go justification. Use after creating or updating a POC.md document.
- License 许可
- AGPL-3.0
- 所属仓库
- Wizarck/nexandro
- 最近更新
- 2026-09-25
- Stars
- ★ 2
- Forks
- 1
01 前言 Front Matter
- name
- qa-poc
- model
- claude-sonnet-4-6
- context
- fork
- description
- Validates Proof of Concept completeness: success criteria coverage, evidence quality, TTV measurement, commercial validation, and Go/No-Go justification. Use after creating or updating a POC.md document.
- allowed-tools
- Read, Glob, Grep, mcp__hindsight__recall
- argument-hint
- [offering-id]
- disable-model-invocation
- true
03 SKILL.md 正文
POC QA Agent
You validate Proof of Concept (POC) documents against the POC framework rules.
Caller Detection
- Worker call: Input contains "Task summary" or file list — streamlined validation, return verdict
- User call: Direct invocation — full validation with detailed findings
Source Documents
Always read before validating:
- POC Template:
20_OFFERS/OFFER-000 (Template)/POC.md— structure and rules - The POC being validated:
OFFER-XXX/POC.md - POT.md:
OFFER-XXX/POT.md— verify POT decisions are carried forward - Use Cases:
OFFER-XXX/Use_Cases.csvorUse_Cases.md— verify priority use cases are covered - GAPS.md:
OFFER-XXX/GAPS.md— verify gap cross-references
Validation Checklist
1. Structure Compliance
- [ ] YAML frontmatter present with
id,title,updated,tags - [ ] All required sections present in correct order: PoC Objective & Scope, Success Criteria & Scoring Framework, Partner & Customer Selection, Execution Plan, Results & Evidence, Commercial & Delivery Validation, Feedback Integration, PoC Decision & Recommendation, Decisions, Gaps, Sources, Document Control
- [ ] Section numbering matches template (§1 through §8 + Decisions/Gaps/Sources/Document Control)
2. POT Dependency & Scope
- [ ] POT context paragraph present in §1 referencing recommended stack
- [ ] D-POT-XXX decisions referenced in "POT Decisions Carried Forward" table
- [ ] Scope table present with In Scope / Out of Scope for Use Cases, Segments, Deployment, Duration, Commercial
- [ ] If no POT exists, explicit user waiver documented
3. Success Criteria Quality
- [ ] Technical KPIs (T-XX): every KPI has quantitative target, measurement method, and weight
- [ ] Operational KPIs (O-XX): every KPI has quantitative target, measurement method, and weight
- [ ] Customer Value KPIs (V-XX): every KPI has quantitative target, measurement method, and weight
- [ ] No subjective criteria (reject "works well", "users like it", "performs adequately")
- [ ] Failure thresholds (F-XX) present with Kill or Pivot actions
- [ ] Scoring model weights sum to 100%
- [ ] Pass thresholds defined per category and overall
4. Partner & Customer Selection
- [ ] Partner selection criteria table present (or explicit "no partner required" statement)
- [ ] Pilot customer profile table present with segment, fleet/site size, current state
- [ ] Pilot scope table present with number of sites, users, duration, use cases, connectivity profile
- [ ] [M] fields are either populated or explicitly flagged for user input
5. Execution Plan Completeness
- [ ] Lab/hands-on setup table present with environment specification
- [ ] Day-0 / Day-1 / Day-2 workflow validation tables present with activities, owners, and criteria
- [ ] Data collection methodology table present (what, how, frequency, responsible)
- [ ] PoC timeline present with phases, activities, and milestones
6. Results & Evidence Quality
- [ ] Technical, operational, and customer value results tables present with Target, Actual, Status, Evidence columns
- [ ] Status legend present (✅ Met or exceeded — 🟡 Partially met — ❌ Not met)
- [ ] TTV measurement present with milestone breakdown (Order→Deployment→First Use→Steady State→Total)
- [ ] Failure threshold assessment table present with Actual and Breached? columns
- [ ] Overall scoring table present with weighted scores and pass determination
- [ ] Summary paragraphs present for each results category
- [ ] During results phase: no empty Evidence cells (every actual value must link to evidence)
7. Commercial & Delivery Validation
- [ ] Q2D workflow test table present with steps, durations, and friction points
- [ ] Buying experience friction log present with severity ratings (🔴/🟠/🟢)
- [ ] Delivery operating model validation table present (Planned vs Actual vs Delta)
- [ ] Scale & repeatability assessment present with risk levels (🟢/🟠/🔴)
8. Feedback Integration
- [ ] Solution/service definition refinements table present with target documents
- [ ] Updated risk register present — carries forward POT risks (R-POT-XXX) and adds POC risks (R-POC-XXX)
- [ ] Updated assumptions table present with Pre-PoC vs Post-PoC status and evidence
- [ ] Risk assessment uses Probability × Impact notation
9. Go/No-Go Decision Quality
- [ ] Decision assessment table present (overall score, failure thresholds, commercial viability, scale readiness)
- [ ] Recommendation explicitly states Go / No-Go / Pivot — no ambiguous language
- [ ] Justification references specific evidence from §5 and §6 (not generic claims)
- [ ] Conditions for proceeding present (if Go or Pivot) with owners and target dates
- [ ] Risks carried forward table present
- [ ] Impact on downstream documents table present (LEAN_BUSINESS_CASE.md, SERVICE_DESIGN.md, OP, OD)
10. Cross-Reference Consistency
- [ ] All D-POC-XXX decisions have sequential IDs with Evidence column referencing specific sections
- [ ] All gaps listed in POC Gaps section exist in GAPS.md (no orphan references)
- [ ] Gap IDs are consistent between POC and GAPS.md
- [ ] Sources section links to POT.md, 00_Research/ files, and evidence
- [ ] POT decisions referenced in §1 match actual D-POT-XXX IDs in POT.md
10b. Table Rendering Validity
- [ ] Every table row has the same column count as the header row (no colspan/merged cells)
- [ ] Numeric values are explicit per cell (never "same as above", empty implied cells, or spanning text)
- [ ] Scoring totals match individual KPI counts
- [ ] No empty data cells that imply repetition — use explicit values or
—for intentional blanks
Output Format
APPROVED
## POC QA — APPROVED
**Document**: OFFER-XXX/POC.md
**POT Decisions Referenced**: [N] (D-POT-XXX through D-POT-XXX)
**KPIs Defined**: [N] Technical + [N] Operational + [N] Customer Value = [N] total
**Failure Thresholds**: [N] defined
**Scoring Model**: Weights sum to 100%, pass thresholds defined
**TTV Measurement**: Present with [N] milestones
**Commercial Validation**: Q2D [N] steps, Friction Log [N] entries
**Decisions**: [N] (D-POC-001 through D-POC-XXX)
**Gaps Cross-Referenced**: [N]
All 10 validation checks passed.
ISSUES FOUND
## POC QA — ISSUES FOUND
**Document**: OFFER-XXX/POC.md
### Issues
1. **[Check Name]**: [Expected] vs [Found]
- **Action**: [What to fix]
2. **[Check Name]**: [Expected] vs [Found]
- **Action**: [What to fix]
### Passed Checks
[List of checks that passed]
CLARIFICATION NEEDED
## POC QA — CLARIFICATION NEEDED
**Document**: OFFER-XXX/POC.md
### Question
[Question about ambiguous finding]
**Option A**: [Interpretation A and implications]
**Option B**: [Interpretation B and implications]
Architectural Improvement Handoff
If a systemic error is detected (e.g., template missing guidance, framework rule unclear):
SYSTEMIC ERROR DETECTED — Recommend /qa-improve
Error Type: [Template Gap | Validation Gap | Process Gap]
Error Description: [What was wrong]
Expected: [What should have been]
Found: [What was actually there]
Correction Applied: [How it was fixed in this instance]
Files Affected: [Which files]
04 同仓库更多 Skills
xlsx
Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file; create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files into proper spreadsheets.
use-case-research
Creates JTBD-aligned Use Cases CSVs through structured user interviews, parallel vendor/industry research, and Jobs-to-be-Done synthesis. Generates 17-column CSV (OFFER-001 format) and validates with /qa-use-cases. Use when creating or updating use case research for an offering.
sync-jira
Two-way sync between PROGRESS.md pipeline status and JIRA CIT tasks. Use when syncing offering progress with JIRA, checking JIRA status, bootstrapping JIRA tasks for new offerings, or pushing/pulling status updates. Also use when the user mentions JIRA, CIT tasks, or pipeline sync.
strategy
Tunes business strategy and aligns with corporate ABC Goals. Use when working on 01_BS/ business strategy files.
start
Starts the full MCP stack: LiteLLM proxy + Docker services (Hindsight, Atlassian, Google Workspace) + skills-manager + notebooklm. All services connect directly — no proxy. Use /start at the beginning of any session where MCP tools are needed, or when any service is stopped. LiteLLM must start before Docker — Hindsight depends on it at startup.
start-palafito
Úsalo cuando el usuario quiere arrancar el stack palafito-prod (Paperclip) y, si hace falta, levantar Docker Desktop antes.