paqad-ai
Version:
Spec-driven development framework — AI agents that think before they type
20 lines (16 loc) • 1.7 kB
Markdown
# Site-map helper skills
verified: 2026-11-03
The helper skills that carry the `site-map` workflow's detail, in the order they run. Each keeps its own procedure; the verbs in the Procedure stay the source of proof.
| Helper | Job |
| ----------------------- | ----------------------------------------------------------------------------------- |
| `site-map-readiness` | grades whether the project is mappable and routes the unmappable case; always first |
| `surface-extraction` | runs the extractors, then normalizes and fingerprints the raw surfaces |
| `surface-modeling` | names, types and attributes the surfaces to modules, with entry and exit marks |
| `transition-tracing` | evidences where navigation goes and what triggers it, with the proving file:line |
| `guard-inference` | evidences the guards on surfaces and transitions, and how each is satisfied |
| `site-map-assembly` | drives the compile verb that folds the layers into `app-map.yaml` |
| `map-verification` | confirms or refutes the claims the deterministic checks could not settle |
| `site-map-gap-analysis` | turns graph invariants and verdicts into `SM-*` findings and a gap report |
| `journey-synthesis` | proposes capped journeys over the verified map, all marked proposed |
| `finding-normalizer` | keeps the finding entries and their vocabulary, including the `SM-*` categories |
| `site-map-retest` | runs the re-check for the Retest mode |