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/data-wise/craft/plangit clone --depth 1 https://github.com/Data-Wise/craftWhat 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 | $0.00035 | $0.01272 |
| Opus 5 | $0.00017 | $0.00636 |
| Sonnet 5 | $0.00007 | $0.00254 |
| Haiku 4.5 | $0.00003 | $0.00127 |
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 yesterday.
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 — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Plan
Single entry point for craft's planning tiers (brainstorm → spec → strategy →
artifact). Decides which tier to invoke via deterministic repo-state detection —
never phrase/keyword classification. That distinction is the entire reason this
command exists safely: phrase classification is the exact mechanism that caused the
project-planner ↔ plan-orchestrator collision (C1,
SPEC-planning-refactor-2026-06-22.md) — a second phrase classifier one layer up
would relocate that risk, not eliminate it. See
GRILL-planning-refactor-a2-2026-07-04.md
G-2 for the full reasoning.
Scope
This command is a router only — it invokes exactly one of craft's existing
planning surfaces and then stops. It never re-implements brainstorm/grill/
plan-orchestrator/project-planner logic itself, and it never invokes /craft:do or
/craft:orch directly: the seam is one-directional (plan produces an
artifact; do/orchestrate consume it in a later, separate invocation — see
SPEC-planning-refactor-2026-06-22.md §4).
Tier Detection (deterministic)
Step 1: Resolve topic
Explicit topic argument, or infer in order from: .STATUS current-task field →
current git branch name → most recent commit subject. This is the same inference
order plan-orchestrator's own "Inputs" section already documents — reused here, not
reinvented.
If no topic can be resolved (no argument, no .STATUS task, an unhelpful branch name
like dev/main), ask via AskUserQuestion rather than guessing.
Step 2: Check repo state for that topic
spec=$(find docs/specs -iname "SPEC-*${topic}*.md" 2>/dev/null | grep -v _archive | head -1)
grill=$(find docs/specs -iname "GRILL-*${topic}*.md" 2>/dev/null | head -1)
orch=$(find . -maxdepth 1 -iname "ORCHESTRATE-*${topic}*.md" 2>/dev/null | head -1)
| Repo state | Route to | Tier |
|---|---|---|
| No SPEC found | /craft:brainstorm (renamed from workflow:brainstorm, v4) |
1 — divergent ideation → SPEC |
| SPEC exists, no GRILL | Offer /craft:grill (ask, don't force — a well-scoped, low-ambiguity spec doesn't strictly need one) |
2 — convergent interrogation |
| GRILL exists (or user declined it), no ORCHESTRATE | plan-orchestrator skill |
4 — artifact generation |
| SPEC + GRILL + ORCHESTRATE all exist | project-planner skill (strategy advice — the only case needing interpretation, since there's no further artifact-presence state left to detect against) |
3 — strategy |
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.
- yesterday First seen · 103 lines · 35 tokens per session scan A 280c8f510208
plan is a command published in the GitHub repository Data-Wise/craft (4 stars, last pushed 16d ago), licensed MIT. It adds 35 tokens to every session and 1,272 once invoked, about $0.0002 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 2026-08-31.
Other commands, from other repositories
readme-audit
Audit a README's reading flow by treating it as a conversion funnel. Every section is evaluated against a target reader and a terminal action. Produces a structured audit blueprint that feeds directly into /readme-restructure.
readme-restructure
Execute a /readme-audit blueprint. Rewrites the README to the evaluator → conversion → links structure and creates linked docs by extracting and reorganising existing content.
build-fix
Fix build and type failures with minimal diffs.
checkpoint
Record a verified checkpoint before the next phase.
doctor
Check the install surface for missing files and invalid manifests.
feature-dev
Drive a feature from plan to implementation to review.