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/spree/spree/update-plangit clone --depth 1 https://github.com/spree/spreeWhat 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 | $0.00535 |
| Opus 5 | $0.00000 | $0.00267 |
| Sonnet 5 | $0.00000 | $0.00107 |
| Haiku 4.5 | $0.00000 | $0.00053 |
Grade A, and why
update-plan 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.
How it starts
The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Update an existing plan document.
Instructions: $ARGUMENTS
- List all plans in
docs/plans/(excluding_template.md) - Identify the relevant plan based on the instructions
- Read the current plan content
- Apply the requested changes while preserving the document structure
- Resolve open questions interactively before finishing — see below
- Update the
Last updateddate - If status changed, update the reference in root
CLAUDE.md
Resolving open questions (required)
Never leave a plan sitting on unresolved open questions after an edit. Before
finishing, check the plan's Open Questions section and use the
AskUserQuestion tool to settle anything the author can decide now — both
questions that were already there and any your changes introduced.
- Ask them together in one call where possible (the tool takes up to 4 questions), so the author sees the whole decision surface at once.
- For each question give 2–4 concrete options, not an open prompt. Lead with
your recommendation and mark it
(Recommended). - Use the
previewfield to show the actual shape of each option — the code, the config, or the schema it implies. - State the cost of each option in its description, not just what it does.
Then fold the answers back into the plan:
- Move each resolved question out of Open Questions into Key Decisions, phrased as a decision with its date and the reasoning that settled it.
- Record any accepted trade-off explicitly, and if it constrains how other code must be written, add it to Constraints on Current Work.
- Leave Open Questions holding only genuinely deferred items, each with a note on what would unblock it.
- If the last blocking question is now settled, move the status off
Draft.
If a question genuinely cannot be answered yet — it depends on work that hasn't landed, or on information nobody has — say so and leave it deferred with the reason. That is a resolution too; silently leaving it unexamined is not.
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 · 49 lines · 0 tokens per session scan A b1ae0f320e03
update-plan is a command published in the GitHub repository spree/spree (15,658 stars, last pushed today), licensed BSD-3-Clause. It costs nothing until one of its globs matches a file; then it loads 535 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
triagem-de-pr
Tria um PR de contribuidor de ponta a ponta — acolhe, mede, reproduz, corrige, responde. Para no merge, que é do mantenedor.
deskcomm-gov-loop
Executa UMA sessão do gov-loop do DeskcommCRM (uma feature de governança, depois morre).
omg-debug
Command "omg-debug" from chenmitchell/omg-payment-skill, covering /omg-debug, 用法, ai 執行流程, 無參數 and mac 參數.
omg-legal
Command "omg-legal" from chenmitchell/omg-payment-skill, covering /omg-legal, 用法, ai 執行流程, 強制免責聲明 and 變數.
omg-bot
Command "omg-bot" from chenmitchell/omg-payment-skill, covering /omg-bot, 用法, ai 執行流程, 硬性規則 and 變數收集.
omg-health
Command "omg-health" from chenmitchell/omg-payment-skill, covering /omg-health, 用法, ai 執行流程, 查詢模式 and 建置模式.