New: 97,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Session Feedback

by gweslab in gweslab/cerf

Propose a durable rule from a failure in this session, for the user to approve.

MITAuto-check passed

Install Session Feedback

skills CLI
$ npx skills add gweslab/cerf --skill session-feedback -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install gweslab/cerf session-feedback --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/gweslab/cerf.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/session-feedback .claude/skills/session-feedback && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
session-feedback
GitHub stars
103
Token cost
~1.3k tokens
SKILL.md length
837 words
Files
1
Skills in repo
11
Repo updated
First seen
Licence
MIT

At a glance

Propose a durable rule from a failure in this session, for the user to approve.

  • Works in 5 steps: When nothing new qualifies, a rule for… → When something qualifies, write each… → Under each bullet add one line, ↳… → …
  • SKILL.md covers The test, Where the entry goes, The highest-value category:… and What does not qualify, plus 4 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Session Feedback is an agent skill from gweslab/cerf. Propose a durable rule from a failure in this session, for the user to approve.

Its SKILL.md is about 1.3k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

The repository describes itself as: Universal Windows CE Emulator - CE Runtime Foundation. The licence is MIT.

Example prompts

  • “/session-feedback”

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

  1. When nothing new qualifies, a rule for the failure already exists and did not stop you, so that rule is broken. Never reply Nothing to add…
  2. When something qualifies, write each entry as one bullet in the voice of its neighbors: a short bold title, an em-dash, one abstract…
  3. Under each bullet add one line, ↳ target: § . This line guides the review and never enters the document.
  4. Propose only what you can defend. One strong entry beats three weak ones.
  5. Stop. Edit nothing. Wait for the user.

What it can do on your machine

Read from SKILL.md and the folder at commit 462ed3d. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Session Feedback loads about 1.3k tokens when it runs. Until then it costs about 24 tokens; SKILL.md has 837 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~24
When it runs · the whole SKILL.md, loaded when a task matches
~1.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from gweslab/cerf at commit 462ed3d, republished under its MIT licence (© gweslab). 837 words, ~1,317 tokens.

Download SKILL.mdSave it as .claude/skills/session-feedback/SKILL.md (or your agent's skills folder).
name
session-feedback
description
Propose a durable rule from a failure in this session, for the user to approve.

Session Feedback

A future agent starts with no memory of this session. When that agent walks into the same wall you hit today, an entry in the durable documents is what stops it. That is the only reason to add one.

Keep the bar high. The documents work because each line earned its place, and every weak line makes the next agent read more to learn less.

The test

Propose an entry only when all of these are true:

  • A future agent repeats the failure. You hit a wall today, and nothing in the current documents stops the next agent from hitting it.
  • The documents do not say it. Not in CLAUDE.md, not in any agent_docs/ page, not in user memory, not even approximately.
  • You confirmed it. You reached the situation and saw the outcome. A hunch does not qualify.
  • It stays true. The entry reads the same in a year. A fact about one function, one bug, or today's tree does not.
  • It saves hours. Minutes do not justify a line that every future agent must read.

Where the entry goes

The entry is aboutTarget
The project - rules, architecture, subsystems, workflowagent_docs/<page> or CLAUDE.md
This machine or this user - local paths, host tooling, personal preferenceUser memory

Project knowledge never goes into user memory. CLAUDE.md § Documentation Targets names this skill and redirects such an entry to the matching agent_docs/ page.

The highest-value category: subsystem map gaps

Ask this first: did agent_docs/subsystems.md mislead you, omit something, or cost you time to find a subsystem that is basic project knowledge? More agent hours die here than anywhere else. Two triggers, and both usually qualify:

  • You hunted for a subsystem that already existed. One line in subsystems.md would have ended the search. That line is missing.
  • You built or reshaped a subsystem. A new agent expects it in basic project knowledge, and subsystems.md does not mention it. Add what it is, where it lives, and what it owns, in the voice of its neighbors.

File paths and subsystem names belong in these entries, because they are the navigation target. Offsets, log lines, and function internals still do not.

What does not qualify

  • Session trivia: today's bug in X, the missing null check in Z.
  • Anything that rots: IDA offsets, log excerpts, file:line pointers, function names, commit hashes, struct layouts. These live in code comments or investigation documents.
  • A restatement of a rule the documents already carry.
  • Speculation. When you would hedge it with "I think" or "seems like", drop it.
  • Padding written because you feel you owe the user an entry.
Show full SKILL.md (407 more words)Show less

Style

Write the principle, not the case study. The rule carries the meaning alone, and the thing that triggered it today disappears.

  • Wrong: "imx51_nfc.cpp returns 0 from its default case, so the guest read garbage for two hours before anyone saw it."

  • Right: "A register switch whose default returns a value hides an unmodelled path. An unimplemented register must fatal, so the gap surfaces at the first touch."

  • Wrong: "The Jornada 720 hangs at boot until SA-1111 register 0x1800 reads back 0x40."

  • Right: "A guest that spins forever on one peripheral register waits for a bit the emulation never sets. Find that bit in the datasheet before you change anything on the guest side."

  • Wrong: "cerf.log line 'EmulatedMemory::Translate unmapped 0x4A000000' means the INTC is missing."

  • Right: "An unmapped-MMIO halt names the peripheral block by its address. Decode that address against the SoC memory map before you suspect the driver."

Do not use offsets, log lines, filenames, or the words "today I found". The claim must read the same five years from now.

Protocol

  1. When nothing new qualifies, a rule for the failure already exists and did not stop you, so that rule is broken. Never reply Nothing to add this session. Reply that the existing rule needs an improvement. Then give it: the document and section, the line as it stands, and the replacement text.
  2. When something qualifies, write each entry as one bullet in the voice of its neighbors: a short bold title, an em-dash, one abstract sentence.
  3. Under each bullet add one line, ↳ target: <document> § <section>. This line guides the review and never enters the document.
  4. Propose only what you can defend. One strong entry beats three weak ones.
  5. Stop. Edit nothing. Wait for the user.

After the review

  • When the user approves, add the approved entries to their targets in the matching style. Drop the ↳ target line.
  • When the user approves some, add those and drop the rest without comment.
  • When the user rejects or rewrites an entry, take that as final and do not pitch it again.
  • When the user says nothing conclusive, drop all of it.

Commit

Once the edits land on disk, run Skill(leak) and Skill(simple-english) over the diff of the edited documents. Fix what they flag. Only then ask in one line whether to commit. Then stop and wait.

When the user approves, invoke Skill(commit) and follow it. That skill owns staging and message style.

© gweslab, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .claude/skills/session-feedback of gweslab/cerf.

Open the folder on GitHubat commit 462ed3d

Compare with similar skills

Session Feedback next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Session Feedback compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Session Feedback this skillgweslab/cerf103—~1.3kAutomated safety check: PassMIT
Rules Distillationaffaan-m/ECC274k2 repos~2.3kAutomated safety check: PassMIT
Geo Proposalsickn33/agentic-awesome-skills47k1 repos~3.2kAutomated safety check: NotesMIT
Feedbackcodewhale-hq/Codewhale41k—~272Automated safety check: PassMIT
Codex Rules Referencecode-yeongyu/oh-my-openagent70k—~269Automated safety check: PassCustom licence
Better Proposals AutomationComposioHQ/awesome-claude-skills77k3 repos~764Automated safety check: PassNone

Similar skills

  • Rules Distillation

    affaan-m/ECC

    Scans installed skills for principles that recur across them and proposes rule-file changes: append, revise, add a section, create a file or leave as covered.

    274k GitHub starsUsed in 2 repos~2.3k tokens
    Agent WorkflowsAuto-check passed

  • Geo Proposal

    sickn33/agentic-awesome-skills

    Auto-generate a professional, client-ready GEO service proposal from audit data.

    47k GitHub starsUsed in 1 repo~3.2k tokens
    Sales & SupportAuto-check: notes

  • Feedback

    codewhale-hq/Codewhale

    Report a Codewhale bug or idea as a GitHub issue. An agent skill from codewhale-hq/Codewhale.

    41k GitHub stars~272 tokensUpdated today
    Auto-check passed

  • Codex Rules Reference

    code-yeongyu/oh-my-openagent

    Explains how the Codex Rules plugin injects project instructions and file-specific rules into a session, which rule files it reads and which settings control it.

    70k GitHub stars~269 tokensUpdated today
    Agent WorkflowsAuto-check passed

  • Better Proposals Automation

    ComposioHQ/awesome-claude-skills

    Automate Better Proposals tasks via Rube MCP (Composio). An agent skill from ComposioHQ/awesome-claude-skills.

    77k GitHub starsUsed in 3 repos~764 tokens
    Productivity & AutomationAuto-check passed

  • Contract And Proposal Writer

    alirezarezvani/claude-skills

    Generate professional, jurisdiction-aware business documents: freelance contracts, project proposals, SOWs, NDAs, and MSAs.

    28k GitHub starsUsed in 2 repos~3.4k tokens
    Legal & ComplianceAuto-check passed

More from gweslab/cerf

All 11 skills in this repo

  • Cerf

    gweslab/cerf

    List the project skills and offer the environment doctor. An agent skill from gweslab/cerf.

    103 GitHub stars~1.2k tokensUpdated yesterday
    Auto-check passed

  • Changelog Append

    gweslab/cerf

    Add a changelog entry for a change that was just made (only user triggered, no agent self-invocation).

    103 GitHub stars~864 tokensUpdated yesterday
    Auto-check passed

  • Commit

    gweslab/cerf

    Create a git commit with a short message that describes the diff.

    103 GitHub stars~1.2k tokensUpdated yesterday
    Auto-check passed

  • Start Board Implementation

    gweslab/cerf

    Start the bring-up of a new board or ROM in CERF. An agent skill from gweslab/cerf.

    103 GitHub stars~3.2k tokensUpdated yesterday
    Auto-check passed

  • Tracking

    gweslab/cerf

    Manage the cross-session tracking document with restore, create, update, or compact (only user triggered, no agent self-invocation).

    103 GitHub stars~22k tokensUpdated yesterday
    Auto-check passed

  • Verify

    gweslab/cerf

    Spawn a hostile reviewer that checks a claim or a diff against the project rules.

    103 GitHub stars~5.4k tokensUpdated yesterday
    Auto-check passed

Questions about Session Feedback

What does Session Feedback do?

Propose a durable rule from a failure in this session, for the user to approve. Session Feedback is an agent skill from gweslab/cerf. Propose a durable rule from a failure in this session, for the user to approve.

How do I install Session Feedback in Claude Code?

Run `npx skills add gweslab/cerf --skill session-feedback -a claude-code`. Or copy the skill folder (.claude/skills/session-feedback in gweslab/cerf) into .claude/skills/session-feedback in your project. Claude Code loads it when a task matches its description.

How do I install Session Feedback in Codex?

Run `npx skills add gweslab/cerf --skill session-feedback -a codex`. Or copy the skill folder (.claude/skills/session-feedback in gweslab/cerf) into .agents/skills/session-feedback in your project. Codex loads it when a task matches its description.

Can I use Session Feedback in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add gweslab/cerf --skill session-feedback -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/session-feedback, .gemini/skills/session-feedback, .github/skills/session-feedback and .opencode/skills/session-feedback in your project.

What does Session Feedback need to run?

SKILL.md names no scripts, command-line tools or credentials: Session Feedback is instructions for the agent only.

Does Session Feedback access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Session Feedback safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Session Feedback use?

Session Feedback is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Session Feedback use?

About 1.3k tokens (SKILL.md is roughly 5.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Session Feedback?

Skills that share tags, products or a category with Session Feedback: Rules Distillation (affaan-m/ECC, 274k stars), Geo Proposal (sickn33/agentic-awesome-skills, 47k stars), Feedback (codewhale-hq/Codewhale, 41k stars) and Codex Rules Reference (code-yeongyu/oh-my-openagent, 70k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Session Feedback?

gweslab (a GitHub organization) maintains it in gweslab/cerf, which has 103 GitHub stars. The repository holds 11 skills in this directory. The repository was last updated on October 7, 2026.

Source: gweslab/cerf on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.