infoq-agent-control-plane
Steal InfoQ Sep 8 2026 control-plane methods, not the webinar or paid SaaS. Use when the user pastes an InfoQ newsletter, mentions Gisting, KYAML, Copilot Review billing, Foundry/HCP control planes, Airbnb auth, DoorDash Flux, or Workload Identity. Fail closed. No Harness/Ori/Copilot-review spend.
- License 许可
- MIT
- 最近更新
- 2026-11-09
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- infoq-agent-control-plane
- description
- Steal InfoQ Sep 8 2026 control-plane methods, not the webinar or paid SaaS. Use when the user pastes an InfoQ newsletter, mentions Gisting, KYAML, Copilot Review billing, Foundry/HCP control planes, Airbnb auth, DoorDash Flux, or Workload Identity. Fail closed. No Harness/Ori/Copilot-review spend.
03 SKILL.md 正文
InfoQ agent control plane (steal the method)
InfoQ 2026-09-08: agents can open PRs, but production still needs tests, security, approvals, and rollback. Shopify Gisting, context engineering, billed Copilot review, Foundry/HCP routers, Workload Identity, and "skip peer review on low-risk PRs" are the transferable bits.
Do not attend the Harness webinar as the deliverable. Do not buy Moderne, Tiger Data, Copilot Code Review, Ori, or Foundry Model Router.
Does it help?
| Surface | Use the paid product? | Use instead |
|---|---|---|
| Hermes Desktop routing | No Foundry/OpenRouter primary | hermes_desktop_router_policy.py + this SPOF check |
| Context dumps / Gisting | No Redis/Shopify tokens | gist_context keeps in/out + two ACs |
| Copilot review (billed, 2d lag) | Never | Existing CI Claude review |
| Native Android/iOS diffs | No AI-only merge | require_device_e2e |
| Scripts/skills-only diffs | Yes, AI approve after tests | allow_ai_approve |
| GCP SA JSON in repo | Never | GitHub OIDC |
Before claiming a delivery path is allowed
python3 scripts/infoq_agent_control_plane.py \
--primary-model hermes-main \
--fallbacks nous-deepseek,glm-5.3 \
--in-scope 'pin Desktop tool turns' \
--out-scope 'ori login' \
--acs 'tests pass|CLI fail-closed' \
--service local_pytest \
--paths 'scripts/infoq_agent_control_plane.py,scripts/tests/test_infoq_agent_control_plane.py' \
--credential github_oidc
Allow only when every JSON ok is true (process exit 0).
High-ROI only
- Independent fallbacks (not the same model, not 3B).
- Compress context to in/out + two ACs. Drop newsletter dumps.
- Block billed-per-review and metered model routers.
- Scripts-only PRs may use existing CI AI review; native diffs need device evidence.
- Prefer OIDC over long-lived keys.
04 同仓库更多 Skills
writing-plans
Use when you have a spec or requirements for a multi-step task, before touching code
workflow-economics-lite
Escape AI pilot purgatory: require completion-loop workflows with one owner, baseline, economic metric, system-of-record integration, and audit trail. Block orphan chatbots and interest-as-ROI claims.
verification-before-completion
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always
value-center-lite
Apply InfoQ Rohrer value-center / VSM five questions (agency + coherence). Use when planning work, shipping releases, or auditing agent autonomy claims. Fail closed on product-only slogans without WQTU/paywall value.
using-superpowers
Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions
using-git-worktrees
Use when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback