agentleFS
Fable-ForgeFable-Forge/fableforge-agent-skills/skills/reporting-discipline/SKILL.md

reporting-discipline

Skill1 starsChanged 53 days agoMITView on GitHub

Use when reporting results, summarizing long logs, documents, tool outputs, agent work, code review findings, debugging evidence, or any task where raw text could bloat the conversation context.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 8 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this order unless the user asks for another format:

    line 21
  2. Distinctive:

    Do not paste huge logs, raw documents, command output, diffs, or agent transcripts by default.

    line 31
  3. Distinctive:

    If raw material is useful, save it to an appropriate file and link or cite that file instead of flooding chat.

    line 33
  4. Distinctive:

    When the user explicitly asks for full raw output and it is large, provide a concise warning and prefer a file artifact.

    line 34
  5. Distinctive:

    Keep progress updates to one or two sentences unless a decision or blocker needs more context.

    line 35
  6. Distinctive:

    Do not claim success from memory or vibes.

    line 40
  7. Distinctive:

    If verification was partial, say exactly what was and was not checked.

    line 41
  8. Distinctive:

    Stop and compress before replying if you are about to include:

    line 46

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Reporting Discipline

全局协作约定

  • 编码、审查、重构或修 bug 时,可显式触发 karpathy-guidelines:先说明假设和取舍,保持改动外科式,定义可验证成功标准。
  • 汇报长日志、长文档、工具输出、agent 工作或调试证据时,可显式触发 reporting-discipline:先给结论、关键证据、风险和下一步,不默认粘贴大段原文。
  • 常用触发语:use karpathy-guidelines and reporting-discipline,或“用 karpathy 打底,并按 reporting discipline 汇报”。

Core Principle

Report the useful signal, not the raw mass. Preserve evidence without dumping it into chat.

Default Report Shape

Use this order unless the user asks for another format:

  1. Conclusion or current status.
  2. Key evidence, with short excerpts only when they matter.
  3. Blockers, uncertainty, or risks.
  4. Next action or verified completion.
  5. Links or file paths for bulky raw material.

Context Budget Rules

  • Do not paste huge logs, raw documents, command output, diffs, or agent transcripts by default.
  • Summarize first; include only the smallest excerpt needed to justify the claim.
  • If raw material is useful, save it to an appropriate file and link or cite that file instead of flooding chat.
  • When the user explicitly asks for full raw output and it is large, provide a concise warning and prefer a file artifact.
  • Keep progress updates to one or two sentences unless a decision or blocker needs more context.
  • In final answers, report what changed, what was verified, and what remains uncertain.

Evidence Standards

  • Do not claim success from memory or vibes. Name the verification that supports it.
  • If verification was partial, say exactly what was and was not checked.
  • For multi-agent or tool-heavy work, summarize the independently verified result rather than relaying every intermediate message.

Red Flags

Stop and compress before replying if you are about to include:

  • More than one screen of raw output.
  • Repeated log lines or stack frames.
  • Full files when a path plus summary is enough.
  • Long reasoning history instead of the outcome and evidence.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.