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/robconery/champion/setupgit clone --depth 1 https://github.com/robconery/championWhat 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 | $0.00043 | $0.02901 |
| Opus 5 | $0.00022 | $0.01451 |
| Sonnet 5 | $0.00009 | $0.00580 |
| Haiku 4.5 | $0.00004 | $0.00290 |
Grade B, and why
setup 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 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
3. **Offer to smooth the road.** Every council command reads and writes `~/.champion/`, which means permission prompts on every run. Offer (one AskUserQuestion) to add `Read(~/.champion/**)` and `Write(~/.champion/**)` t How it starts
The opening of the file, as written. The whole thing — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Plugin note: this command ships in the
championplugin. It spawns no agents; you run the interview yourself, in the main conversation.
/setup — the interview
You are about to build this person's accountability mirror: a council that is also them, future them reaching back to pull present them toward what they're trying to do. The council can't fight for someone it doesn't know. This interview is how it learns who it fights for.
The full philosophy lives in ${CLAUDE_PLUGIN_ROOT}/interview/INTERVIEW.md — read it before you begin. This command is the compressed, structured implementation: 20–30 minutes, one question at a time, multiple choice wherever a choice fits, every question carrying its reason. The frame is professional growth, not therapy: you will touch personal territory, but only because personal patterns drive professional choices. Every tender thing you ask about earns its place by connecting to the work.
What setup writes, and where, is specified in ${CLAUDE_PLUGIN_ROOT}/references/dossier-schema.md. Read that too. The dossier root is ~/.champion/dossier/.
Before the first question
- Check for an existing dossier. If
~/.champion/dossier/interview-log.mdexists:- If
PROFILE.mdalso exists, setup previously completed. Offer (via AskUserQuestion): revise one section, re-run from scratch, or leave it alone. Honor$ARGUMENTSif it already says which. - If the log exists but
PROFILE.mddoesn't, a previous run was interrupted. Offer to resume from where it stopped (the log shows the last phase reached) or start over.
- If
- Say what this is, briefly. Three or four sentences: this is a one-time interview, about 20–30 minutes, that builds the dossier every council command reads. Some questions are a tap on a choice; a handful need a real answer in their own words. They can pause any time — progress is saved — and they can revise any section later with
/champion:setup revise. - Make the privacy promise, and mean it. Everything they say is written to
~/.champion/on this machine and nowhere else. This plugin never sends it anywhere; deleting that folder erases the mirror completely. If they are on a shared or employer-managed machine, say so plainly and let them decide before proceeding. If you cannot honestly guarantee privacy in their environment, stop.
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 · 96 lines · 43 tokens per session scan B 7ae6313098ee
setup is a command published in the GitHub repository robconery/champion (5 stars, last pushed 29d ago), licensed MIT. It adds 43 tokens to every session and 2,901 once invoked, about $0.0002 per session on Opus 5. 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
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.