← ClaudeAtlas

plan-assumption-verifierlisted

Power skill that cross-verifies each planning assumption against concrete repo evidence, converting unresolved ones into raised blocking questions.
JeelVankhede/agentsmyth · ★ 0 · AI & Automation · score 72
Install: claude install-skill JeelVankhede/agentsmyth
# Plan Assumption Verifier ## Purpose For each assumption (`A` ID) carried into Plan from the brief, confirm it is backed by concrete repo evidence. Where evidence is absent, convert it into a raised blocking question (`Q` ID, mirrored into `orchestration.blockers`) rather than letting it remain a silent, unverified assumption. Prevents a plan from being built on an assumption nobody actually checked. This is a power skill, not a lifecycle phase. It is active/gate-bound: invoked during Plan's workflow, and its result gates Plan's Exit Gate. ## Invocation Context Use this skill when: - Plan is processing the brief's `A` IDs while writing its own artifact - a Plan revision introduces a new assumption Do not invoke it for a plan with zero carried-over assumptions — nothing to verify. ## What To Load **Foundation** (confirm in context; load if not already present): - Root `AGENTS.md` - `workflow/router.md` - `workflow/lifecycle.md` - `workflow/rules.md` **Minimum for invocation**: - This file - `references/output-schema.md` **Before starting work**: - `references/evidence-standards.md` — what counts as sufficient evidence for an assumption - The brief's Requirement Manifest `A` and `Q` sections - `workflow/config/repo-profile.yaml` — for path/contract-based assumption checks ## Inputs - The brief's `A` IDs. - Repo state relevant to each assumption's claim. ## Refusal / Stop Conditions Stop and raise a `Q` ID instead of silently proceeding when: - an assumption can