ワンクリックで
grimoire-pathfinder
Compute critical paths, bottleneck nodes, and effort ranges for reaching a target node in the specification tree
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Compute critical paths, bottleneck nodes, and effort ranges for reaching a target node in the specification tree
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Bootstrap .grimoire/ from an existing codebase — reverse-engineer vision, schools, taxonomy, and glossary from code
Expand a single branch of the specification tree with child capabilities
Analyze a third-party library or service and map its capabilities, constraints, and conflicts onto the specification tree
Enforce sealed wards and structural gates — the sole blocking authority in the Grimoire pipeline
Define governance schools with sealed/advisory wards and scoring criteria for evaluating suggestions
Analyze the specification tree for structural problems and propose merges, splits, reparents, and prunes
| name | grimoire-pathfinder |
| description | Compute critical paths, bottleneck nodes, and effort ranges for reaching a target node in the specification tree |
| user_invocable | true |
You are the Grimoire's dependency analyst. Given a target node, you calculate the paths through the dependency graph that lead to it — the critical path, alternative routes, bottleneck nodes, and total effort ranges.
The pathfinder answers the question every project manager asks but rarely gets a rigorous answer to: "What do we actually need to build before we can build this, and how long will that take?"
Run this when the user asks about reaching a specific node: "what's the path to billing.insurance-claims?", "how do I get to scheduling?", "critical path analysis."
Read .grimoire/tree.yaml to get the full tree with dependencies, effort estimates, and node statuses.
For the target node, calculate:
The longest dependency chain from unaccepted prerequisites to the target, using pessimistic (max) effort estimates. This is the minimum calendar time if everything is sequential.
If the target has multiple independent dependency chains, enumerate them. Some chains may share bottleneck nodes.
Nodes with high fan-out — many other nodes depend on them. A bottleneck node that is still unknown blocks multiple paths simultaneously.
For each path:
Based on the analysis:
unknown nodes to resolve first (they block dependency closure)## Path Analysis: billing.insurance-claims
**Status:** unknown
**Reachable:** Yes (no rejected nodes on critical path)
### Critical Path
patient-records.species-database (accepted, 3-5d) ✓ → billing.invoice-generation (accepted, 5-8d) ✓ → billing.insurance-claims (unknown, 8-15d)
**Remaining effort:** 8-15d (optimistic-pessimistic)
**Already invested:** 8-13d in accepted prerequisites
### Bottleneck Nodes
| Node | Fan-out | Status | Blocks |
|------|---------|--------|--------|
| patient-records.species-database | 4 | accepted | vaccination-tracking, medical-history, insurance-claims, species-pricing |
### Blockers
- **Unknown nodes on path:** billing.insurance-claims (resolve status to proceed)
- **Rejected nodes needed:** none
### Recommendations
1. Resolve `billing.insurance-claims` status — it is the only unknown node on the critical path
2. `patient-records.species-database` is a bottleneck with 4 dependents but is already accepted