メニュー
Record agent decisions for post-hoc analysis and audit. Use when reviewing agent behavior patterns or tracking decision trends over time.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Generate and maintain AI_STANCE.md for any uFawkes* repo. Use when onboarding a repo to the uFawkesAI suite, when reviewing AI policy currency, or when a new AI tool is being adopted. Implements DORA AI Capability 1.
Ensure graphify corpus is current and complete before each agent session. Define minimum AI-accessible documentation per repo. Use at session startup to verify internal context is available to AI tools. Implements DORA AI Capability 3.
15-minute JTBD + acceptance criterion exercise. Use before any spec session. Produces a discovery-brief.md that anchors the entire increment to a real user need and a measurable DORA outcome. Implements DORA AI Capability 6.
Enforce minimum documentation standard across uFawkes* repos. Use when auditing a repo before release, onboarding a new repo to the suite, or implementing roadmap item 0.6 (cross-repo README links). Directly supports DORA AI Capability 3 and the 2023 DORA finding that documentation quality amplifies all technical capabilities.
Create and manage DOJO content for platform engineering education and knowledge sharing.
Maintain AI_STANCE.md as a living document. Use quarterly or when a new AI tool is adopted. Handles review triggers, the three-bucket update process, socialization checklist, and cross-repo consistency. Companion to ai-stance skill.
| name | agent-decision-logging |
| description | Record agent decisions for post-hoc analysis and audit. Use when reviewing agent behavior patterns or tracking decision trends over time. |
| license | MIT |
| compatibility | Claude Code, GitHub Copilot, OpenCode, Cursor, Codex, Gemini CLI |
| metadata | {"author":"paruff","suite":"uFawkesAI"} |
Load trigger:
"load agent-decision-logging skill"> DORA: Cap 6 (Reliability) Token cost: Low
Record what agents decided and why. Enables post-hoc analysis of agent behavior patterns, decision trends, and identification of systematic biases.
agent.decision.madeEmit when the agent produces its final decision.
Attributes:
| Attribute | Type | Description |
|---|---|---|
decision | string | Final decision (PASS/FAIL/BLOCKED/APPROVE/REQUEST_CHANGES/ESCALATE) |
blocker_count | integer | Number of CRITICAL or HIGH severity findings |
finding_count | integer | Total number of findings |
agent.name | string | Agent that made this decision |
session_id | string | Unique invocation identifier |
| Trend | Meaning | Action |
|---|---|---|
| Increasing FAIL rate | Agents are stricter or quality is dropping | Investigate root cause |
| Decreasing FAIL rate | Agents are looser or quality is improving | Review if standards are slipping |
| Constant APPROVE rate | Agent may not be thorough | Review findings per invocation |
The decision and blockers fields in the Phase 0 invocation log are the source of truth.
The agent.decision.made span is the OTEL representation of those fields.