Borrowing it
Nothing to install: this file belongs to incubateur-ademe/benefriches. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/incubateur-ademe/benefriches/main/.agents/skills/reflect/SKILL.mdgit clone --depth 1 https://github.com/incubateur-ademe/benefrichesWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/incubateur-ademe/benefriches/reflect)<a href="https://agentmods.dev/skills/incubateur-ademe/benefriches/reflect"><img src="https://agentmods.dev/badge/skills/incubateur-ademe/benefriches/reflect/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/incubateur-ademe/benefriches/reflect"><img src="https://agentmods.dev/badge/skills/incubateur-ademe/benefriches/reflect.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00085 | $0.01613 |
| Opus 5.5 | $0.00034 | $0.00645 |
| Sonnet 5.5 | $0.00017 | $0.00323 |
| Haiku 4.5 | $0.00009 | $0.00161 |
Grade B, and why
reflect scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
description: Reflect on the current session to find mistakes, user corrections, ambiguous instructions, and failures — then propose durable root-cause fixes to the project's instruction docs (AGENTS.md, .claude/rules, .a How it starts
The opening of the file, as written. The whole thing — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Reflect - Session Self-Improvement
Reflect on the current session and propose durable, root-cause fixes to the project's
instruction surface (AGENTS.md, .claude/rules/*, .agents/skills/*) so the mistakes,
misunderstandings, or ambiguous instructions that occurred don't recur.
Bias toward the opposite of a point fix: when something went wrong, the interesting output is not "fix this file" but "which instruction doc was ambiguous / missing / mis-scoped such that a careful agent got it wrong — and how should that doc be reworded?"
This is deliberately different from revise-claude-md:
- Signal philosophy: anchored on externally-observable failure signals (corrections, doc-misreads, failing commands, loops) — never open-ended "what could I have done better" introspection. Unaided self-critique of otherwise-correct work degrades quality; don't do it.
- Target surface: the full instruction surface (
AGENTS.md+ rules + skills), not justAGENTS.md. - Fix type: biased toward editing the ambiguous line that misled you, not just appending newly-discovered commands/gotchas.
Step 1 — Detect
Scan the in-context conversation for observable failure/friction signals only:
- Explicit user correction/redirect — "no, do X instead", "actually…", "that's not what I meant", "why did you…", reverting a decision. Highest-precision signal.
- Ambiguous instruction that misled — a rule/AGENTS.md/skill line was followed literally but wrongly; the guidance admitted a wrong reading.
- Tool/test/build/lint failure caused by you — non-zero exits, failing typecheck/lint/ test/build traceable to your own change.
- Loop / repeated retry — you tried the same thing multiple times, thrashed, or went in circles before it worked.
- Revert —
git revert/git checkout --/ undo of your own work.
For each signal found, extract an evidence bundle: the exact quoted turn(s) or command(s) that prove it happened.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed · -1 tokens per session scan A → B e660531057aa
- 1mo ago Changed · -1 tokens per session scan C → A 4f7eb6be090e
- 2mo ago First seen · 136 lines · 87 tokens per session scan C daafe858fc58
reflect is a skill published in the GitHub repository incubateur-ademe/benefriches (45 stars, last pushed today), licensed MIT. It adds 85 tokens to every session and 1,613 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
autoreview
Structured code review through the shared OpenClaw agent-skills installation.
omh-code-review
This is an OMH code-review workflow skill, projected for Agent Skills hosts (Claude Code, Codex, Cursor, opencode, OpenClaw, pi).
revdiff-plan
Review the last Codex assistant message (plan, analysis, or proposal) with inline annotations in a TUI overlay. Extracts the most recent response from Codex rollout files and opens it in revdiff for review and annotation. Activates on "revdiff-plan", "review plan with revdiff", "annotate plan", "review last response"…
explore-unknowns
Guide the user through a quadrant walk that maps the unknowns of a task — open by listing the known knowns, then work through known unknowns, unknown knowns, and unknown unknowns one stage at a time, ending with a complete four-quadrant map in the user's hands. Use when a request is ambiguous or underspecified, the…
re0-merge
Review and land an external contribution the way this suite does: gate it against the thesis, land it with the author's credit intact, complete a new skill rather than merging it raw, then approve, credit, and explain before closing. Use when reviewing a pull request, as any collaborator or maintainer, not only the…
asplos-writing-style
Use when drafting or revising ASPLOS prose — engineering the first two pages that rapid review actually reads, stating the cross-layer insight as one quotable sentence, structuring per-layer mechanism sections, writing bounded claims, and fitting everything into 11 self-contained pages that include figures and…