本文へ移動
cccskills
無料GitHub で公開

fix-p0-issues

Use when the user asks to fix P0s, address critical issues, or work on priorities from the weekly product briefing. Don't use for non-P0 issues or local bug fixes.

インストール方法を見る

含まれるファイル(2)

  • SKILL.md5.1 KB
  • agents/openai.yaml182 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Fix P0 issues

Contract

FieldBound contract
TriggerUser asks to fix P0s, address critical issues, or work on priorities from the weekly product briefing.
AuthorityRemote: spawns ODIN cloud coding agents for triaged, actionable P0s; requires explicit human invocation. No agent is spawned without explicit user confirmation.
Side effectTriages P0s, spawns high-confidence agents, and reports run IDs and monitoring commands.
DoneActionable P0s have run IDs; skipped items are documented; no agent runs without user confirmation.

Inputs

  • The repository root, resolved via git rev-parse --show-toplevel.
  • The most recent weekly product briefing under reports/weekly_product_briefings/*.md.
  • Optional: GitHub issue numbers embedded in briefing P0 sections as #NNNN.

Procedure

  1. Resolve the repository root. Find the most recent briefing by sorting filenames, which contain YYYY-MM-DD dates. Do not use ls -t because filesystem modification times are unreliable in freshly cloned repos.
    REPO_ROOT=$(git rev-parse --show-toplevel)
    ls "$REPO_ROOT/reports/weekly_product_briefings/"*.md 2>/dev/null | sort | tail -1
    
    Done when: the most recent briefing is identified by path.
  2. Parse the briefing for sections starting with #### P0:. For each P0, extract the title (text after #### P0:), GitHub issue numbers (#NNNN patterns), the problem description paragraph(s), and any engineering-alignment notes. Done when: every P0 section is parsed with title, issue numbers, description, and alignment notes.
  3. When a P0 includes issue numbers, fetch the GitHub issue details (gh issue view <issue_number> --repo <org>/<repo>). Then search the codebase using keywords from the issue. Document relevant file paths, related code or tests, and dependencies. Done when: every P0 with issue numbers has its issue fetched and codebase searched.
  4. Triage each P0 into one category and record the decision:
    • Inactionable Complaint: no specific description, reproduction steps, or fix indication. Skip; note in the report.
    • Bug Fix: clearly scoped, has reproduction steps or a clear error, narrow scope, identifiable root cause. Decision: Spawn Agent.
    • Feature Request: assess scope (Small 1-2 files, Medium 3-10 files, Large 10+ files or architectural), complexity, and confidence 1-5. Confidence 5 or 4 → Spawn Agent (4 with caveats); 3 or lower → Skip and note. Skip any P0 the briefing notes is already being addressed on an active branch. Do not spawn more than 3 agents at once. Done when: every P0 has a category and decision recorded.
  5. For each P0 marked Spawn Agent, construct a prompt containing the issue number and title, problem description, fetched issue details, relevant file paths, and task (investigate, implement fix, add or update tests, create a PR with a clear description). Present the full spawn plan to the user, including P0 titles, issue numbers, categories, confidence, decisions, and constructed prompts. Obtain explicit confirmation before proceeding. Do not spawn any agent without confirmation. Done when: the spawn plan is presented and explicit confirmation is obtained or refused.
  6. For each confirmed P0, spawn an ODIN cloud coding agent:
    odin agent run-cloud \
      --environment <ODIN_ENVIRONMENT_ID> \
      --prompt "<constructed prompt>"
    
    Use exactly odin to spawn cloud agents. If odin is not available, stop and report the error rather than substituting another binary. Done when: every confirmed P0 has a cloud agent spawned (or odin is unavailable and the skill stops).
  7. Record each spawned agent's run ID and emit the report described in Output. Done when: every spawned agent's run ID is recorded and the report is emitted.

Failure and recovery

  • No briefing found: stop and report that no weekly product briefing exists under reports/weekly_product_briefings/; do not spawn agents.
  • odin unavailable: stop and report the error; do not substitute another binary or spawn method.
  • GitHub issue fetch fails: continue with briefing-only context for that P0; note the fetch failure in the report.
  • User declines confirmation: do not spawn any agent; report the declined plan as the terminal result.
  • Partial-result rule: agents already spawned remain spawned; report their run IDs. No rollback of remote runs is performed by this skill.

Output

Use this report form:

=== P0 Fix Agents Spawned ===

P0: <Title>
- GitHub Issues: #NNNN, #NNNN
- Category: <Bug Fix | Feature Request | Inactionable Complaint>
- Confidence: <1-5> (Feature Requests only)
- Decision: <Spawn Agent | Skip>
- Run ID: <run-id>
- Monitor: odin run get <run-id>

...

Skipped:
- <Title>: <reason>

To check all runs:
odin run list --output-format text

Skipped P0s (inactionable, low-confidence, or already addressed) are listed with their reason. No agent runs without user confirmation.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Use when explaining System V AMD64, ARM AAPCS, RISC-V psABI, stack frames, variadic calls, or FFI register rules. Not for the Rust FFI binding layer: use rust-ffi.

日本語の概要は準備中です。原文の説明を表示しています。

OutlineDriven/odin-claude-plugin372026年9月29日 更新

Use when configuring ADC sampling time, DMA-driven ADC, calibration, or DAC channel setup on bare-metal MCUs. Not for the DMA stream itself: use dma-baremetal.

日本語の概要は準備中です。原文の説明を表示しています。

OutlineDriven/odin-claude-plugin372026年9月29日 更新

af-xdp

無料

Use when creating AF_XDP sockets, configuring UMEM and XSK rings, writing an XDP redirect program, or choosing copy versus zero-copy mode. Not for full kernel bypass: use dpdk.

日本語の概要は準備中です。原文の説明を表示しています。

OutlineDriven/odin-claude-plugin372026年9月29日 更新

Use when a completed session needs an agent-environment retrospective. Not for an engineering retrospective from telemetry: use engineering-retrospective.

日本語の概要は準備中です。原文の説明を表示しています。

OutlineDriven/odin-claude-plugin372026年9月29日 更新

Use when a redacted, trimmed agent transcript must be appended to a GitHub PR or issue body, with human approval and preview. Not for automated or model-initiated insertion.

日本語の概要は準備中です。原文の説明を表示しています。

OutlineDriven/odin-claude-plugin372026年9月29日 更新

agents-md

無料

Use when a repo needs agent setup, AGENTS.md added or made lean, CLAUDE.md audited, or agent instructions scored or pruned. Not for remote, credential, publish, deploy, or irreversible changes.

日本語の概要は準備中です。原文の説明を表示しています。

OutlineDriven/odin-claude-plugin372026年9月29日 更新

OutlineDriven のスキルをすべて見る

このスキルの問題を報告する