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 commands/valpere/chorus/debuggit clone --depth 1 https://github.com/valpere/chorusWhat 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 | $0.00013 | $0.00552 |
| Opus 5 | $0.00006 | $0.00276 |
| Sonnet 5 | $0.00003 | $0.00110 |
| Haiku 4.5 | $0.00001 | $0.00055 |
Grade A, and why
debug 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 yesterday.
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.
What it actually says
Generate parallel root-cause hypotheses from five agents and synthesize a ranked investigation plan.
Raw slash-command arguments:
$ARGUMENTS
Debug focus per agent:
- Claude — application logic, state management, data flow
- Gemini — infrastructure, concurrency, external dependencies, environment
- Codex — edge cases in input handling, type coercion, off-by-one errors
- Cursor — framework, library, and third-party integration issues
- Kilo — naming, types, readability, and long-term maintainability
Execution mode:
- Default: foreground (debugging needs immediate results).
- If
--backgroundis in the arguments, run in background mode. - If
--jsonis in the arguments, the companion emits{"command":"debug","results":[{name,output,error,exitCode}]}on stdout instead of delimited text. Warnings still go to stderr. - Strip
--background,--wait, and--jsonbefore passing to companion.
Pre-flight: The companion automatically checks which agents are installed. Missing agents are reported in stdout with install instructions. Debugging proceeds with whatever agents are available — at least 2 are required; if fewer are installed the companion exits non-zero and you should tell the user to install more agents.
Foreground execution:
node "$CLAUDE_PLUGIN_ROOT/scripts/companion.mjs" debug $ARGUMENTS
Background execution:
Bash({
command: `node "$CLAUDE_PLUGIN_ROOT/scripts/companion.mjs" debug $ARGUMENTS`,
description: "Parallel debug",
run_in_background: true
})
Tell the user: "Parallel debug started in the background. You'll be notified when it completes."
After the companion exits, synthesize an investigation plan:
## Hypothesis Pool
| # | Hypothesis | Proposed by | Likelihood |
|---|-----------|-------------|------------|
| 1 | … | Claude | High |
| 2 | … | Gemini + Codex | Medium |
| … |
## Investigation Plan
1. Check X first (quick, rules out hypotheses 1 and 3)
2. …
## Most Likely Root Cause
<1–2 sentences based on the aggregate council input.>
Rank hypotheses by likelihood, surfacing ones flagged by multiple agents first.
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.
- yesterday First seen · 67 lines · 13 tokens per session scan A 50c693d87ee6
debug is a command published in the GitHub repository valpere/chorus (10 stars, last pushed 3mo ago), licensed MIT. It adds 13 tokens to every session and 552 once invoked, about $0.0001 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
update-ai-versions
Look up the latest published versions of the AI coding assistants pinned in the Dockerfile and update the ARG lines in place.
yapper-help
The /yapper-help command displays full documentation for the Yapper plugin, including all modes, commands, behaviours, configuration options, and usage examples. This command is itself rendered in Professor mode by default, because a help document that is not thoroughly elaborated would be somewhat contrary to the…
yapper-review
The /yapper-review command performs a thorough, multi-line pull request review. Rather than posting terse inline comments like "this is wrong" or "use X instead," yapper-review produces fully-explained feedback that describes the issue, the reasoning behind the concern, the potential impact, and a concrete suggested…
yapper
The /yapper command activates the Yapper verbosity system. Once active, all subsequent responses in the session will be substantially expanded, elaborated, and enriched with definitions, transitions, and restatements.
yapper-commit
The /yapper-commit command generates a verbose, fully-explained Git commit message for the current change or a described change. The resulting commit message explains not just what changed, but why it changed, what problem it solves, and how the solution works — providing maximum context for future readers of the git…
mem-prune
Removes old low-importance memories.