Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/robinslange/learning-loop/rewritenpx skills add robinslange/learning-loop --skill rewritegit clone --depth 1 https://github.com/robinslange/learning-loopWhat 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 | $0.00067 | $0.02494 |
| Opus 5 | $0.00034 | $0.01247 |
| Sonnet 5 | $0.00013 | $0.00499 |
| Haiku 4.5 | $0.00007 | $0.00249 |
Grade A, and why
rewrite scanned grade A with 0 findings 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 2d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 208 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Rewrite: Context Gene Editing
When a belief turns out to be wrong (or refined), the wrong version sits in three places: the vault as durable notes, the auto-memory as preferences/feedback, and the episodic history as past conversations the LLM may resurface. This skill edits all three coherently.
The Katsuno-Mendelzon split:
- Vault and auto-memory use revision semantics: actively rewritten or archived. The world model changes.
- Episodic memory is append-only history: update semantics. We do not edit past conversations; we annotate them via the supersessions table so future retrievals carry the correction inline.
When to Use
/learning-loop:rewrite "old pattern" "new pattern": full form, with reason inferred/learning-loop:rewrite "old pattern" "new pattern" "reason": explicit reason for the supersession record/learning-loop:rewrite: no args, infer the change from recent conversation context
Process
Phase 1: Frame the Change
If args are provided, parse old_pattern, new_pattern, optional reason.
If no args, read the conversation. The user just learned something that contradicts a prior belief. Identify:
- The OLD pattern (the belief being retracted/refined)
- The NEW pattern (the replacement)
- The REASON (what evidence forced the change)
If you cannot find a clear correction in the conversation, tell the user and stop. Ask them to provide old and new explicitly.
Confirm the framing in one line before proceeding:
Rewriting: "<old>" → "<new>" (reason: <reason>)
Phase 2: Hit Map (parallel search across all stores)
Search every store for the OLD pattern. Run all four searches in parallel (single message, multiple Bash calls):
-
Vault: semantic + keyword:
node ${CLAUDE_PLUGIN_ROOT}/scripts/vault-search.mjs search "<old pattern>" --rerank -
Vault: wiki-link/title hits: Use
Globfor filename matches:**/*<key-noun>*.mdin{{VAULT}}/ -
Auto-memory:
Grepthe project's~/.claude/projects/*/memory/directory for substrings of the old pattern. ReadMEMORY.mdand any matching files.
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.
- 2d ago First seen · 208 lines · 67 tokens per session scan A 57a97eb2d54c
rewrite is a skill published in the GitHub repository robinslange/learning-loop (11 stars, last pushed 10d ago), licensed Apache-2.0. It adds 67 tokens to every session and 2,494 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…