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/chankov/agent-fleet/af-compoundgit clone --depth 1 https://github.com/chankov/agent-fleetWhat 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 | $0.00015 | $0.00385 |
| Opus 5 | $0.00008 | $0.00192 |
| Sonnet 5 | $0.00003 | $0.00077 |
| Haiku 4.5 | $0.00002 | $0.00038 |
Grade A, and why
af-compound 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 3d 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
Invoke the compound-learning skill via the skill tool.
Capture this session's lessons into the project's own rules (HOW) and docs (WHAT/WHY), optionally focused on $ARGUMENTS:
- Resolve the targets: read
.ai/agent-fleet-overrides.md— the## agent-hub(legacy## agent-team) section'srules:anddocs:keys name the rule folders and documentation entry points. If absent, locate an existing rules/docs tree and confirm it with the user; never invent a new tree. - Gather the evidence: this conversation (user corrections, decisions, rejected approaches), the session's
git diff/git log, and any review findings produced during the session. - Extract at most 5 candidate lessons — each one imperative sentence plus a one-line Why (the failure it prevents) and a one-line Evidence (what happened this session). Classify each as rule, doc, or neither; the default verdict is neither.
- Dedupe index-first: read the rules tree's
README.md/index.mdmanifest, grep for existing coverage, and prefer sharpening an existing rule in place over adding a new one. - Propose the surviving lessons — target file, Why, Evidence, and the exact diff — and wait for approval. The user filters preferences-of-the-moment from durable policy.
- Apply the approved diffs as minimal edits to existing files (at most 1 new file, registered in the tree's index), verify links and indexes still resolve, and report the changes file by file.
Run this at the end of a session, before the context is gone. If nothing rises to a lesson, say so and stop — compounding a smooth session produces filler.
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.
- 3d ago First seen · 17 lines · 15 tokens per session scan A 96601d867a7d
af-compound is a command published in the GitHub repository chankov/agent-fleet (10 stars, last pushed 7d ago), licensed MIT. It adds 15 tokens to every session and 385 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
git
Git operations with intelligent commit messages and workflow optimization.
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.