agentleFS
Works-HQWorks-HQ/ai-brain-starter-pack/skills/reflect/SKILL.md

reflect

Skill3 starsChanged 3 months agoMITView on GitHub

Capture corrections so they do not recur and run a weekly review of how the system is working. Use after a correction, when the workflow feels wrong, or once a week. Triggers on "reflect", "correct once, never again", "weekly reflection", and "run the weekly reflect review".

172 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use poteto-help, setup-pstack and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • ianclearySkill

    poteto-help

    Guides users through pstack setup, /poteto-mode, and picking the skill, playbook, or principle for a task. Type /poteto-help with a question.

    Also in 14 projects that use this

  • bnemaSkill

    setup-pstack

    Configure which Pi models pstack uses per role. Detects available provider/model specs and writes pstack's user configuration. Use for /setup-pstack, "configure pstack models", or changing pstack model choices.

    Also in 24 projects that use this

  • eryzerzSkill

    make-bot-ui

    Use when building a custom UI (page, dashboard, buttons) that should wake a Grok Bot over a webhook, when the user must provide a webhook sender key, or when exposing that UI on Tailscale.

    Also in 19 projects that use this

  • ianclearySkill

    principle-explain-the-number

    Apply before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.

    Also in 26 projects that use this

  • jaydubya818Skill

    automate-me

    Use when the user wants working preferences captured in a reusable mode skill.

    Also in 39 projects that use this

  • jaydubya818Skill

    poteto-mode

    Use when non-trivial work should follow the smallest rigorous Jstack workflow.

    Also in 27 projects that use this

  • +6 more from jaydubya818

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 12 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 163 edited versions we read keep these.

  • 12

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Done when: every finding the user selected is written to its target file, and the user has confirmed what the system is now smarter about.

    line 10
  2. Very specific:

    Company mode: use this when operating-system/ exists with operating-system/GLOBAL_TASKS.md.

    line 16
  3. Very specific:

    Personal mode: use this when root-level TASKS.md, debrief-history.log, or about-me.md shows the personal structure.

    line 17
  4. Distinctive:

    Both detected: ask which mode to use. Write nothing until the user chooses.

    line 18
  5. Distinctive:

    Do not create a structure just to make the skill run.

    line 21
  6. Distinctive:

    Use this after the user corrects an assumption, style, fact, or process.

    line 25
  7. Distinctive:

    Wait. Persist only the selected findings.

    line 31
  8. Distinctive:

    Use this when the user asks for a weekly reflection.

    line 35
  9. Distinctive:

    Prefer editing an existing line over adding another rule.

    line 48
  10. Distinctive:

    Do not edit a brain or skill file during this run.

    line 55
  11. Distinctive:

    Never write before the user selects what to persist.

    line 59
  12. Distinctive:

    Read a target before editing it.

    line 60
  13. Distinctive:

    Keep decisions, tasks, and status in debrief.

    line 62
  14. Distinctive:

    After writing, confirm what the system is now smarter about and name the file changed.

    line 65

“Common” is measured: kept by at least 6 in 10 of the 163 edited versions of this file we could read. “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

Reflect

Improve how the system works, not the work itself. This skill has two explicit triggers: correction-capture and weekly review.

Done when: every finding the user selected is written to its target file, and the user has confirmed what the system is now smarter about.

Detect the structure first

Inspect the current folder before planning any write.

  1. Company mode: use this when operating-system/ exists with operating-system/GLOBAL_TASKS.md.
  2. Personal mode: use this when root-level TASKS.md, debrief-history.log, or about-me.md shows the personal structure.
  3. Both detected: ask which mode to use. Write nothing until the user chooses.
  4. Neither detected: say that this folder does not contain a recognised personal or company structure. Point the user to START-HERE. Write nothing.

Do not create a structure just to make the skill run.

Trigger A: correction-capture

Use this after the user corrects an assumption, style, fact, or process. The aim is correct once, never again.

  1. Identify direct corrections, explicit preferences, repeated friction, and weak inferences.
  2. Sort them as high confidence, medium confidence, or observation only.
  3. Propose a specific durable instruction for each useful finding.
  4. Show the numbered findings and their target files. Ask which numbers to persist.
  5. Wait. Persist only the selected findings.

Trigger B: weekly system review

Use this when the user asks for a weekly reflection.

  • In personal mode, read debrief-history.log, completed or changed work in TASKS.md, about-me.md, and relevant CLAUDE.md files.
  • In company mode, read operating-system/DEBRIEF_LOG.md, the Done section of operating-system/GLOBAL_TASKS.md, the previous entries in operating-system/REFLECTION_LOG.md, and operating-system/current-context.md.

Assess what worked, what did not, repeated patterns, stalled work, missing context, and concrete changes for the coming week. Show the proposed reflection and every intended write, then ask which findings to persist. Wait for the answer.

Route approved findings

Personal mode

  • A global preference about voice, format, or working style goes in about-me.md.
  • A correction or rule specific to one company or area goes in that area's CLAUDE.md.
  • Prefer editing an existing line over adding another rule.
  • A weekly review must result in at least one user-selected, concrete improvement to about-me.md or a relevant CLAUDE.md. If the user selects none, write nothing.

Company mode

  • Record both correction-capture and weekly-review findings as a dated entry in operating-system/REFLECTION_LOG.md.
  • Include the correction, the evidence, the proposed durable change, and its target file.
  • Do not edit a brain or skill file during this run. Leave approved changes as explicit actions in the reflection log so they can be reviewed separately.

Approval and quality rules

  • Never write before the user selects what to persist.
  • Read a target before editing it. Do not duplicate an existing instruction.
  • Be specific. "Use short sentences and no jargon" is stronger than "be casual".
  • Keep decisions, tasks, and status in debrief. Keep corrections, preferences, and system patterns here.
  • Be direct and actionable. Do not pad the reflection with reassurance.

After writing, confirm what the system is now smarter about and name the file changed.

Scaling to per-project folders

Company mode deliberately keeps one top-level operating-system/REFLECTION_LOG.md. If separate teams later need their own review histories, add per-project logs and update this routing rule after the shared weekly review has proved useful.

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.