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/rjmurillo/ai-agentsWrote 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/rjmurillo/ai-agents/plan)<a href="https://agentmods.dev/commands/rjmurillo/ai-agents/plan"><img src="https://agentmods.dev/badge/commands/rjmurillo/ai-agents/plan/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/rjmurillo/ai-agents/plan"><img src="https://agentmods.dev/badge/commands/rjmurillo/ai-agents/plan.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 | $0.00023 | $0.00714 |
| Opus 5.5 | $0.00009 | $0.00286 |
| Sonnet 5.5 | $0.00005 | $0.00143 |
| Haiku 4.5 | $0.00002 | $0.00071 |
Grade A, and why
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 2mo 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 — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
@CLAUDE.md
Plan: $ARGUMENTS
If $ARGUMENTS is empty, check for recent /spec output in the conversation. If none found, ask the user what to plan.
If you arrived here without a spec, run the front-gate first
Planning a spec that was never gated manufactures work. If there is no /spec output for this work (no requirement, no design, no testable acceptance criteria), do not decompose it into milestones yet. Run /spec first: it applies the front-gate (the six forcing questions, the front-gate-before-pipeline pattern) and confirms a named blocked user, a documented status quo, and a concrete observation before any downstream step runs. Return here once the spec exists. Skip this only when the user explicitly asks to plan an ungated idea and accepts that trade-off.
Process
- Read the spec or issue
- Map sub-problems to existing code (what already exists? use Grep/Glob to verify)
- Task(subagent_type="milestone-planner"): You are a project planner. Break the spec into milestones with clear exit criteria. Each milestone is independently shippable. Sequence by dependencies. Flag parallel opportunities.
- Task(subagent_type="task-decomposer"): You are a work breakdown specialist. Decompose each milestone into atomic tasks. Each task is independently verifiable with a clear done definition. Size by complexity (S/M/L), not time.
- Invoke Skill(skill="execution-plans") to persist the plan as a versioned artifact.
- Task(subagent_type="analyst"): You are a risk analyst. Run a pre-mortem on this plan. What fails first? What dependencies are fragile? What assumptions are untested?
- Task(subagent_type="critic"): You are a plan reviewer. Validate: is scope complete? Can tasks execute in the stated sequence? Are estimates credible? Is anything missing?
Evaluation Axes
- Scope integrity - Nothing unnecessary, nothing missing
- Dependency ordering - Can tasks execute in the stated sequence?
- Risk coverage - All P0 risks have mitigations
- Estimate confidence - Complexity-based sizing (S/M/L), not time-based
- Reversibility - Which steps are hard to undo?
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.
- 2mo ago First seen · 50 lines · 23 tokens per session scan A 93701a56a97c
plan is a command published in the GitHub repository rjmurillo/ai-agents (49 stars, last pushed today), licensed MIT. It adds 23 tokens to every session and 714 once invoked, about $0.0001 per session on Opus 5.5. 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
cs-internal-comms
Internal-only change-management comms using ADKAR (Prosci) + Kotter's 8-step. NOT marketing (external) and NOT executive narrative strategy. Direct invocation of the internal-comms skill.
cs-knowledge-ops
Company SOP + runbook authoring with 5W2H completeness checks. NOT personal PKM (that's llm-wiki). NOT engineering-specific runbooks. Direct invocation of the knowledge-ops skill.
cs-process-map
Map an internal business process (BPMN-style swim lanes), measure cycle time, and detect bottlenecks where work spends most of its time waiting. Direct invocation of the process-mapper skill.
research-verify
Verify existing research findings against independent primary sources. Upgrades confidence from 'sources agree' to 'independently verified.'.
plan-change
Interactive change-planning session to prepare context for /write-spec.
write-plan
Generate or update implementation plan.