← ClaudeAtlas

vault-circuitolisted

Trigger: internal step-9 runbook invoked by /report and /reporte-lote after a circuit's report renders. Projects the circuit's 3 validated narrative JSONs into reports/vault/{circuito}.md and chains an isolated /graphify reports/vault --update against reports/vault's own graph (never the project-root graph). Not intended for direct end-user invocation.
amalvarezme/chec-local-uiti-vano-interpreter · ★ 0 · Web & Frontend · score 65
Install: claude install-skill amalvarezme/chec-local-uiti-vano-interpreter
## Overview `vault-circuito` is step 9 of `/report`'s run sequence: a pure post-render projection that turns a circuit's already-validated `historical.out.json`/`inference.out.json`/`expert-alignment.out.json` into one upserted Spanish markdown note at `reports/vault/{circuito}.md`, then chains the real `/graphify reports/vault --update` slash-command — scoped to an isolated graph rooted at `reports/vault/graphify-out/graph.json`, never the project-root `graphify-out/graph.json` a whole-project `/graphify .` run produces — so the vault stays incrementally indexed without ever mixing with or corrupting the unrelated project-wide code/docs graph. It makes NO LLM calls of its own — `vault_note_contract.py` performs pure file I/O — and it never touches `report_pipeline.py`'s HTML critical path (steps 2-8 already succeeded and returned their result before this step ever runs). ## When to Use Loaded automatically by [`report/SKILL.md`](../report/SKILL.md)'s step 9, immediately after step 8 has already reported the rendered HTML path — both for a standalone `/report` run and for each circuit inside a `/reporte-lote` batch (which runs `report/SKILL.md`'s steps 2-9 per circuit, per [`reporte-lote/SKILL.md`](../reporte-lote/SKILL.md)). Not designed for direct end-user invocation — there is no standalone `/vault-circuito` slash-command trigger. It may be run manually for debugging via the CLI verb directly (see Related artifacts). ## Argument contract Takes exactly one argument: `c