Archon is a workflow engine for AI coding agents that turns development processes into YAML-defined sequences with phases, validation gates, and artifacts. Developers use it to run repeatable processes such as planning, implementation, testing, code review, and pull-request creation across projects. The catalogue entries provide commands, agents, skills, hooks, instructions, and settings for working with Archon.
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/coleam00/ArchonWrote 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/coleam00/archon/review-synthesize)<a href="https://agentmods.dev/commands/coleam00/archon/review-synthesize"><img src="https://agentmods.dev/badge/commands/coleam00/archon/review-synthesize/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/coleam00/archon/review-synthesize"><img src="https://agentmods.dev/badge/commands/coleam00/archon/review-synthesize.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 | $0.00000 | $0.03010 |
| Opus 5 | $0.00000 | $0.01505 |
| Sonnet 5 | $0.00000 | $0.00602 |
| Haiku 4.5 | $0.00000 | $0.00301 |
Grade A, and why
review-synthesize 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 11d 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 — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Decide the Review
Produce one evidence-based verdict, write the review report humans read, and publish it when the scope is a PR. You are read-only: never modify project files or commit; your only write outside the artifacts directory is the PR comment below. Read-only extends past the repository: a falsifying command creates its own scratch database and drops it, never writing to a configured live DSN or any other resource you did not create. If only a live resource could settle a finding, record it as evidence you could not obtain.
There are two modes in $ARTIFACTS_DIR/review/scope.md:
- Full review: aggregate the independent specialist reports. Connect and prioritize their evidence; do not perform another broad review or invent findings.
- Continuation review: continue from the previous report as the reviewer. Verify its findings, review the correction delta, and decide whether the change converged. Do not treat the latest SHA as a new PR and do not repeat the specialist fan-out in your own head as a checklist.
Read the review state
- Read
$ARTIFACTS_DIR/review/scope.md, and the project'sarchitecture.mdif it has one, then inspect the exact diff scope.md records. Verify claims against the code, never against summaries. Anchor on the accepted work order's stated invariants and on the same risk scaling the lenses use — irreversible or destructive paths, lifecycle ownership, persisted contracts and schemas, credentials and auth boundaries, integration boundaries, concurrency over shared state; a risk this change engages that no lens engaged is incomplete review, not a clean verdict. - In full mode, read every current specialist report present in
$ARTIFACTS_DIR/review/(code.md,seams.md,simplify.md,tests.md,errors.md,docs.md) in full. - In continuation mode, read
$INPUTS.prior_reportin full before judging the delta. Also read$ARTIFACTS_DIR/implementation.mdwhen it exists; it records what the correction claims to have changed and proved. Specialist files beside the report belong to the earlier round and are evidence only through the canonical prior report. Do not count them as freshly rerun lenses. - Read every producer record under
$ARTIFACTS_DIR/discoveries/, when that directory exists. Its absence means no producer recorded a discovery. Each file is independent evidence; never delete or replace these raw files.
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.
- 11d ago First seen · 110 lines · 0 tokens per session scan A faaa2e1de64f
review-synthesize is a command published in the GitHub repository coleam00/Archon (23,418 stars, last pushed today), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 3,010 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 2026-08-30.
Other commands, from other repositories
review-prs
Parallel PR triage: decide merge-worthiness, prepare rebased worktrees, fix blockers, return them for human merge.
fix-issues
Diagnose, reproduce, then fix reproducible open GitHub issues in parallel: one clean worktree/issue; symlink build artifacts to avoid rebuilds.
cleanup
Autonomous cleanup-loop iteration: discover ONE target → complete execution → verify → report. Runs stateless: derive from current tree; assume prior runs left it consistent.
review
Run code review on files or recent changes.
hatch3r-pr-resolve
Read open PR comments, evaluate each against current code via the rigor contract, implement accepted findings, reply inline. Multi-platform.
factory-ticket
Implement exactly one already-claimed Linear ticket in the current worktree.