Agent skill

deliverable-selfloop

Self-loop a deliverable until it is publishable by the pipeline standard: diagnose -> fix -> re-check, and write a PASS/FAIL report. **Trigger**: self loop, self-loop, polish deliverable, quality gate, fix-on-fail, 收敛, 自循环, 质量门. **Use when**: A pipeline has produced a reader-facing deliverable (`output/*.md`) and you want deterministic convergence to PASS. **Skip if**: You are still pre-approval for prose or the upstream evidence/structure artifacts are missing. **Network**: none. **Guardrail**: Do not invent papers/citations/results. Only use in-scope inputs already present in the workspace.

View SKILL.md on GitHub Repository
Stars 377
Forks 25

Install this agent skill to your Project

npx add-skill https://github.com/WILLOSCAR/research-units-pipeline-skills/tree/main/.codex/skills/deliverable-selfloop

SKILL.md

Deliverable Self-Loop (fix-on-fail)

Goal: converge a pipeline deliverable to a stable, reader-facing quality bar.

For the ideation path, the target is now a brainstorm memo bundle:

  • output/REPORT.md
  • output/APPENDIX.md
  • output/REPORT.json
  • plus the supporting trace chain under output/trace/

Load Order

  1. Always read references/overview.md first — defines trace chain, convergence model, and upstream routing.
  2. Read references/quality_checklist.md before running the gate — lists required sections, quality tokens, and templated-language anti-patterns checked by the script.

Inputs

  • output/trace/IDEA_SIGNAL_TABLE.md
  • output/trace/IDEA_DIRECTION_POOL.md
  • output/trace/IDEA_SCREENING_TABLE.md
  • output/trace/IDEA_SHORTLIST.md
  • output/REPORT.md
  • output/APPENDIX.md
  • output/REPORT.json

Output

  • output/DELIVERABLE_SELFLOOP_TODO.md (report-class; always written)

Workflow

  1. Run the gate — run the script (see Script below) on the workspace. It checks trace presence, shortlist size, required sections, appendix quality, and templated language.

  2. Read the TODO report — open output/DELIVERABLE_SELFLOOP_TODO.md. If - Status: PASS, the deliverable is done.

  3. Fix blockers — for each issue listed under ## Remaining blockers:

    • Missing output/trace/IDEA_SIGNAL_TABLE.md → run idea-signal-mapper.
    • Missing output/trace/IDEA_DIRECTION_POOL.md → run idea-direction-generator.
    • Missing output/trace/IDEA_SCREENING_TABLE.md → run idea-screener.
    • Missing output/trace/IDEA_SHORTLIST.md → run idea-shortlist-curator.
    • Missing REPORT sections → expand the report via idea-memo-writer.
    • Thin APPENDIX → enrich with per-paper reading guides.
    • Templated language → rewrite canned phrases to context-specific language.

  4. Rerun until PASS.

Acceptance

  • output/DELIVERABLE_SELFLOOP_TODO.md exists and contains - Status: PASS.
  • All 7 trace/output artifacts present and non-empty.
  • Shortlist size is 3–5.
  • No templated language detected.

Script (optional; deterministic gate)

Quick Start

  • python .codex/skills/deliverable-selfloop/scripts/run.py --workspace workspaces/<ws>

All Options

  • --workspace <dir> (required)
  • --unit-id <U###>
  • --inputs <semicolon-separated>
  • --outputs <semicolon-separated>
  • --checkpoint <C#>

Examples

  • Run the gate on a brainstorm workspace:
    • python .codex/skills/deliverable-selfloop/scripts/run.py --workspace workspaces/brainstorm-llm-agents

Didn't find tool you were looking for?

Be as detailed as possible for better results