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.
git clone --depth 1 https://github.com/jonase47/ccprWrote 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/commands/jonase47/ccpr/cleanup)<a href="https://agentmods.dev/commands/jonase47/ccpr/cleanup"><img src="https://agentmods.dev/badge/commands/jonase47/ccpr/cleanup/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/commands/jonase47/ccpr/cleanup"><img src="https://agentmods.dev/badge/commands/jonase47/ccpr/cleanup.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 | $0.00000 | $0.03603 |
| Opus 5.5 | $0.00000 | $0.01441 |
| Sonnet 5.5 | $0.00000 | $0.00721 |
| Haiku 4.5 | $0.00000 | $0.00360 |
Grade B, and why
cleanup 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 26d 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.
grep -m1 '^HANDOVER_WARN_PCT' ~/.claude/hooks/agent-monitor.py How it starts
The opening of the file, as written. The whole thing — 225 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/cleanup – Doc Hygiene: HANDOVER cap + lint aggregator
Runs a one-shot hygiene pass on a project's docs to keep them lean and
machine-readable: HANDOVER inbox triage and size cap enforcement plus the four
existing lint scripts (memory-lint.sh, phase-docs-lint.sh, doc-volume-check.sh,
manual-lint.sh) bundled into one consolidated drift report. Use this between phases (after a Gate)
or whenever the project-guide warns about HANDOVER size, an unprocessed inbox,
stale memory or doc volume drift.
Argument: $ARGUMENTS = [optional: projectdir]
- Without argument: operates on
$(pwd). - With argument: operates on the given project directory.
Flow
1. HANDOVER inbox triage (confirm before removing)
The HANDOVER carries an append-only inbox (## Open Points, see
templates/HANDOVER_TEMPLATE.md): working agents drop findings there that fall outside their
assignment, when running in the main working tree (a worktree-isolated agent reports the line in
its final message instead), and this step is the one place they are cleared again.
Scan docs/HANDOVER.md for lines matching the marker pattern ^- INBOX [|] — match on the
marker, not the heading, so entries appended under a differently named heading are still found.
Write the separator as the bracket expression [|], never as a bare |: bare, it is alternation in
ERE (grep -E) and in GNU BRE (grep '\|'), where ^- INBOX OR empty matches every line of
the file. [|] is a literal pipe in BRE, ERE and Python re alike. Reference command:
grep -c '^- INBOX [|]' docs/HANDOVER.md
It must return the same count with -E added — if the two differ, the pattern is wrong, not the file.
- No matches → report
inbox: 0 entriesand continue with §2. Do not create the section: a project on an older template stays valid without it. If the## Open Pointsheading is missing entirely, mentiontemplates/HANDOVER_TEMPLATE.mdonce in the report — as a hint, not an action. A freshproject-initHANDOVER is expected to report 0: the template's format example lives inside the section's blockquote (>) and therefore does not match the marker. - Matches → split each line on
|intomarker | date | source | finding | refand present all entries as a table with one proposed target each. If a line yields more than five fields (someone wrote a literal|into the finding text), keep field 1 as the marker, fields 2–3 as date and source, the last field asref, and re-join everything in between as the finding — the reference is always last. Never drop the surplus text and never re-cut the line in the file.
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.
- 26d ago Changed · +1 lines 6cff04c738ab
- 1mo ago Changed · +1 lines 34933410e1fb
- 1mo ago Changed 44fbc177bfd4
- 1mo ago Changed edd7bbef1b68
- 1mo ago Changed · +3 lines 1599cb9511ab
- 1mo ago First seen · 220 lines · 0 tokens per session scan B 8bd16bac7354
cleanup is a command published in the GitHub repository jonase47/ccpr (1 stars, last pushed 26d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 3,603 tokens. 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 2026-08-31.
Other commands, from other repositories
stats
Show real token usage, delegation rate, and actual savings for the current Hydra session (parses Claude Code session JSONL — no AI estimation).
map
View, rebuild, or query the codebase dependency map.
design-intent
Guided UI/visual development with structured exploration, architecture design, and quality review.
ralph
Ralph autonomous development loop with fresh context per iteration.
debug
Systematic debugging with automated investigation and 4-phase methodology. Default: inline evidence gathering and diagnosis. --deep: spawns systematic-debugger agent for full autonomous debugging. Use for errors, stack traces, test failures, or unexpected behavior.
double-check
Comprehensive verification with automated quality gates, confession gathering, and confidence-based issue filtering. Default: self-review with gates. --deep: multi-agent review via code-reviewer agent. Use before committing, after fixing bugs, or to validate feature completeness.