address

A command that updates a written implementation plan after another person or tool has reviewed it. It reads the plan and the review file stored beside it.

In plain words
What is it for?
Use it to address review findings in development plans, check those findings against the actual code, and record any required changes to related plans.
Why use it?
It prevents changes from being made based on an unverified or missing review. It also helps distinguish valid findings from outdated or incorrect comments.

Command

Install

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.

agentmods
npx agentmods add commands/cosmix/loom/address
Clone the repo
git clone --depth 1 https://github.com/cosmix/loom
Per session 8 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 468 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00008 $0.00468
Opus 5 $0.00004 $0.00234
Sonnet 5 $0.00002 $0.00094
Haiku 4.5 $0.00001 $0.00047

Measured 2d ago against content hash 8fa95f6c6614, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

address 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.

commands/address.md · 19 lines

What it actually says

read the review of your plan. It is at codex-<basename of $1> in the SAME directory as the plan (for example, when $1 is doc/plans/PLAN-foo.md the review is doc/plans/codex-PLAN-foo.md). Read $1 (the plan file) alongside it.

⚠️ IF THAT REVIEW FILE DOES NOT EXIST, STOP. It means the reviewer failed to produce one. Reply saying exactly that and change NOTHING — not the plan, not planning-omissions.md. Do NOT substitute a review of your own: this command's only job is to fold in someone else's findings, and an unreviewed edit pass on the plan is worse than no pass at all.

Work through each point in the review. Validate each point against the ACTUAL code and the sibling plans BEFORE accepting it — reviews are sometimes written against a stale draft (line references drift; a point may already be folded in) and are sometimes wrong. If a point is valid, fix the plan to address it. If a point is not valid or already addressed, explain why in your reply — not in the plan.

When an accepted fix touches a contract another plan provides or consumes (symbol names, signatures, file ownership, seams), fold the fix in AND state the required sibling-plan amendment explicitly in the plan as a blocking note — never silently diverge from a sibling.

After folding all fixes, run two closing sweeps: (1) self-consistency — rg each changed claim across the whole plan file and reconcile prose ⇄ YAML (a half-applied correction is worse than none); (2) coverage — every capability the updated prose promises must appear in a stage's artifacts:/wiring:/acceptance; a fix that lands only in prose is invisible to loom's gates.

⚠️ THE PLAN MUST STAY SELF-CONTAINED: fold every accepted fix directly into the plan file, because the executor reads only the plan. Do not move required detail into external notes.

Record any omissions, mistakes, or bad assumptions this review surfaced about your original plan in planning-omissions.md at the root of the repo. If the file is still there, append to it.

Changes

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.

  1. 2d ago First seen · 19 lines · 8 tokens per session scan A 8fa95f6c6614

Subscribe to this mod's changes

address is a command published in the GitHub repository cosmix/loom (54 stars, last pushed 4d ago), licensed MIT. It adds 8 tokens to every session and 468 once invoked, about $0.0000 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 2026-08-30.