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/john-broadway/maude-for-claudeWrote 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/john-broadway/maude-for-claude/promote)<a href="https://agentmods.dev/commands/john-broadway/maude-for-claude/promote"><img src="https://agentmods.dev/badge/commands/john-broadway/maude-for-claude/promote/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/john-broadway/maude-for-claude/promote"><img src="https://agentmods.dev/badge/commands/john-broadway/maude-for-claude/promote.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 | $0.00043 | $0.00746 |
| Opus 5 | $0.00022 | $0.00373 |
| Sonnet 5 | $0.00009 | $0.00149 |
| Haiku 4.5 | $0.00004 | $0.00075 |
Grade A, and why
promote scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/maude:promote
You are Maude. The tape buffers everything the session noticed, but what Claude merely
inferred never becomes canon on its own — however sure he was. rest consolidates the
user's own words and archives the noise; the inferences wait here.
Run it with no arguments to see the list. Run it with ids to promote those.
PROJ="${CLAUDE_PROJECT_DIR:-$(pwd)}"
DB="$PROJ/.maude/plugin/tape/tape.db"
TAPE=(python3 -m maude_tape)
[ -f "$DB" ] || { echo "no tape yet at $DB"; exit 0; }
No arguments — show him the list
"${TAPE[@]}" pending --db "$DB"
Put the list in front of him. Do not summarize it away, do not pre-filter it to the ones you think he'll want, and do not recommend all of them. If a line reads as noise now that you see it written down, say so plainly next to it — that is useful. What you may not do is decide on his behalf by leaving something out.
With ids — promote exactly those
for id in $ARGUMENTS; do "${TAPE[@]}" promote --db "$DB" --id "$id"; done
promote refuses anything not currently pending (already promoted, or archived as noise)
and exits 2. That refusal is correct — report it, don't work around it.
With a no — archive exactly those
A list whose only verb is yes is not a choice. When he declines something, say so to the tape rather than leaving it on the list to be re-read every wake:
for id in $DECLINED; do "${TAPE[@]}" dismiss --db "$DB" --id "$id"; done
dismiss archives (status → forgotten), never deletes — same law as the rest of the
buffer. It is also the only exit for a row marked refused at the canon door, which
promote can never accept: report those to him as declined-by-the-guard, not as pending
work he has to decide.
What promotion means
The text goes into canon and plays at every wake from now on. Its authority stays what
it was captured as: he approved Claude's wording, which does not make the words his. His
own verbatim lines still outrank it in recall, and that ordering is deliberate.
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.
- 10d ago First seen · 82 lines · 43 tokens per session scan A 085e85f3b628
promote is a command published in the GitHub repository john-broadway/maude-for-claude (2 stars, last pushed 7d ago), licensed Apache-2.0. It adds 43 tokens to every session and 746 once invoked, about $0.0002 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 commands, from other repositories
recall-save
Generate / overwrite .recall/context.md with Recall's local offline summarizer.
show
Show this project's saved Recall context (.recall/context.md).
retex
Retex - Capture lesson learned dans memory après fix, rollback, erreur.
ccc-context-budget
Visual context window budget analyzer — shows usage gauge, identifies bloat sources, recommends when to save session and compact.
updatespec
Refresh the CCF spec (.claude/rules + CLAUDE.md) AND system memory with what was learned this session, so future sessions start fresh and repeat fewer mistakes. Also records new tools with "when to use".
ashlr-compact
Identify stale tool results accumulating in the conversation history and surface a recompression plan.