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/lexfrei/ccc/submitnpx skills add lexfrei/ccc --skill submitgit clone --depth 1 https://github.com/lexfrei/cccWhat 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 | $0.00092 | $0.02050 |
| Opus 5 | $0.00046 | $0.01025 |
| Sonnet 5 | $0.00018 | $0.00410 |
| Haiku 4.5 | $0.00009 | $0.00205 |
Grade A, and why
submit 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 — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Every posting of a series — v1 and every reroll — is a new thread. Review feedback is answered in the old thread (lkml:reply); fixed code is posted fresh. The workflow below is ordered as gates: content, recipients, format, timing, send. A failed gate stops the send.
1. Content gates (before formatting anything)
Commit message = WHY. The subject is subsystem: imperative summary at 70-75 characters; the body says why the change is needed and why this way, hard-wrapped at 75 columns. Mechanism walkthroughs, rejected-alternative essays, and reachability analyses do not belong in the commit body — put them under the --- scissors or in the cover letter. A verbose commit body draws "This is very verbose. Is it AI generated?" and a changes-requested; there is a public precedent of exactly that: https://lore.kernel.org/netdev/[email protected]/
WHY-only is necessary but not sufficient — there is also a volume cap. The working measure is maintainer-tip.rst's improved examples: context, problem, solution as separate paragraphs in that order, two to three paragraphs of two to three sentences. Write for a reviewer who has the diff open — anything readable from it gets deleted, and a body that narrates what the code does draws "Blah, blah, blah. Please assume i can read the patch and understand what it does" even when every sentence is a why (precedent, the v2 of that same series: https://lore.kernel.org/netdev/[email protected]/). When in doubt, too short beats too long.
The scissors are not a dumping ground either. Analysis moved below --- still gets read, and a long one draws "condense the above writing in a much shorter text, or point to prior discussion" while burying the one sentence a reviewer actually needs ("Why conditional? Maybe it is buried deep in that wall of text?"). Keep the scissors section to a few sentences and replace analysis with a lore link to the thread where it already happened; a justification a reviewer is likely to ask about belongs in the body, briefly, not below the cut.
Verify claims about sets. Any sentence of the form "all drivers do X", "this lock is only taken once", "no caller passes NULL" is an invitation for a reviewer to find the counterexample. Grep the tree for each such claim before sending, or rewrite it as a statement about the mechanism.
Bug fixes carry a Fixes: tag — 12+ character SHA plus the quoted subject of the broken commit, one line, never wrapped. The commit to name is the oldest one that still reproduces the issue, found with git log -S rather than guessed; when the root cause genuinely predates git history, the canonical tag is Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2"). A fix posted without the tag gets asked for it and rerolled. Reports from other people carry Reported-by: plus a Closes: link to the report on lore.
AI assistance is disclosed per Documentation/process/coding-assistants.rst with the kernel's own tag format, which checkpatch validates:
Assisted-by: LLM
Format is Assisted-by: LLM [tool1] [tool2], the optional brackets naming analysis tools like coccinelle or smatch when they were actually used. The Assisted-By: Name <email> form used in other projects is formally invalid here.
Rerolls answer everything first. A new version goes out only after every comment on the previous one has a reply in the old thread, and never while that discussion is still live. Questions a reviewer had to ask get their answers folded into the commit message itself, not just the thread — per maintainer-netdev.rst, "occasionally the update of the commit message will be the only change in the new version".
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 · 83 lines · 92 tokens per session scan A 676f234522c3
submit is a skill published in the GitHub repository lexfrei/ccc (9 stars, last pushed 2d ago), licensed BSD-3-Clause. It adds 92 tokens to every session and 2,050 once invoked, about $0.0005 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 2026-08-31.
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…