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/fprochazka/claude-code-plugins/write-plangit clone --depth 1 https://github.com/fprochazka/claude-code-pluginsWhat 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 | $0.00018 | $0.02340 |
| Opus 5 | $0.00009 | $0.01170 |
| Sonnet 5 | $0.00004 | $0.00468 |
| Haiku 4.5 | $0.00002 | $0.00234 |
Grade A, and why
write-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 — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Write an Implementation Plan
Write a plan file for the work described below, then hand it off for execution. The plan must be shaped so that (a) the resulting git history is clean per the git-workflow rules, and (b) the implementation can be driven by one persistent implementation subagent that receives the plan file itself, not giant inlined excerpts.
This command is normally run after /sdlc:pre-plan and a design discussion, so most context should already be in hand — don't re-derive what's already been gathered or decided.
Scope
$ARGUMENTS
If the Scope is empty, use the current conversation context (e.g. a pre-plan briefing already produced in this session). If it's unclear what we're planning, ask the user before proceeding.
Step 1 — Load the git-workflow skill
Invoke the git:git-workflow skill first, and apply its rules to shape the plan around the ideal git history rather than the order code happens to get written. The concrete tie-in for this command: each step in the plan maps onto one intended atomic commit (or a small, named group), so the plan's structure already encodes the commit sequence.
Step 2 — Front-load every decision, before plan mode
Do not enter plan mode yet. Plan mode restricts what the permission system lets you run, and this phase may still need real work — subagent explorations, data checks, queries against external systems. Enter plan mode only in Step 3, when the only work left is writing.
The plan you write will be executed with minimal supervision — the user may not be at the keyboard while it runs. So the bar is not "do I know enough to start writing": it is "have I surfaced every decision the user would otherwise be asked mid-implementation". A question asked now costs one AskUserQuestion; the same question at step 7 stalls the whole run.
- Walk the intended steps and hunt for the decisions hiding inside them: user-facing naming, product behavior at edge cases, scope trade-offs, anything irreversible. Ask about all of it now (AskUserQuestion) — the goal is that the executing agent never has to stop for a decision.
- Genuine knowledge gaps in the code → fill with targeted, delegated exploration (subagents, not inline spelunking) — only for the gaps, not a fresh sweep.
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 · 91 lines · 18 tokens per session scan A 81fef51d83e3
write-plan is a command published in the GitHub repository fprochazka/claude-code-plugins (11 stars, last pushed 4d ago), licensed MIT. It adds 18 tokens to every session and 2,340 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 2026-08-30.
Other commands, from other repositories
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.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.