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/cdeust/ai-architect-mcp-codebase/composegit clone --depth 1 https://github.com/cdeust/ai-architect-mcp-codebaseWhat 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 | $0.00000 | $0.00351 |
| Opus 5 | $0.00000 | $0.00176 |
| Sonnet 5 | $0.00000 | $0.00070 |
| Haiku 4.5 | $0.00000 | $0.00035 |
Grade A, and why
compose 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.
What it actually says
Compose Multiple Genius Agents
Load multiple genius agents in sequence, apply each to the problem, and pass output forward.
Instructions
-
Parse $ARGUMENTS: agent names separated by spaces, then
--, then the problem description. Example:/genius-compose curie fermi -- "We see 3x more latency than our model predicts but have no measurement of where it goes" -
Run
tools/genius-invoker.sh compose <agent1> <agent2> ... -- "<problem>"to validate all agents exist and load their content. -
For each agent in sequence: a. Read the full agent file at
agents/genius/<name>.mdb. Apply that agent's complete workflow to the problem (or to the accumulated output from prior agents) c. Produce the output in that agent'soutput-formatd. Label the output clearly: "=== Output from ===" -
Chain rule: each agent after the first receives:
- The original problem description
- The full output from all previous agents
- An instruction to build on, refine, or challenge the prior output using its own method
-
After all agents have run, produce a Synthesis section that:
- Identifies where the agents agree (convergent insights)
- Identifies where they disagree or see different aspects (complementary insights)
- Proposes concrete next actions that combine the strongest outputs
- Notes any gaps that no agent addressed
-
If any agent's workflow calls for artifacts (difficulty books, provenance files, estimates), create them using the corresponding tools.
$ARGUMENTS
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 · 32 lines · 0 tokens per session scan A 3a182ab7199f
compose is a command published in the GitHub repository cdeust/ai-architect-mcp-codebase (4 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 351 tokens. 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
brooks-audit
Run a Brooks-Lint architecture audit.
brooks-health
Run a Brooks-Lint codebase health dashboard across all four dimensions.
daily-standup
Génération Résumé Daily Stand-up.
deploy
Build, test, deploy with staged rollout.
fix-issue
Analyze and fix a reported bug or issue systematically.
simplify
Reduce complexity without changing behavior — code simplification.