From Change Reviewer
The severity taxonomy for change-review findings — how to assign BLOCKING / CONCERN / NOTE so the release gate enforces the right things and never false-blocks, and how to ground invariant/risk reasoning in the explorer MEMORY.md. Use when a reviewer sub-agent or /reviewer:run records a finding.
How this skill is triggered — by the user, by Claude, or both
Slash command
/reviewer:check-invariantsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Severity decides what GATES. The pipeline release gate enforces **0 BLOCKING**; CONCERN is
Severity decides what GATES. The pipeline release gate enforces 0 BLOCKING; CONCERN is recorded but does not block; NOTE is informational and excluded from the tally. Assign deliberately — an inflated BLOCKING blocks a clean change; a buried BLOCKING ships a real breakage.
Ground every judgment in the explorer memory. .claude/explorer/MEMORY.md records the
invariants (" — breaks if: <…>") and the Risk map
("- — — — "). A finding is strongest when it cites the exact
invariant the change violates, with path:line evidence from the diff.
Use for: a dangling caller, a dropped safety contract, or a violated invariant the codebase relies on to stay correct.
.sh that dropped set -uo pipefail, stopped sourcing ../lib/common.sh, or
(re)introduced set -e — the house preamble that prevents fail-open. [R2]Use for: risk-mapped surface and scope discipline.
MEMORY.md Risk map — re-read the recorded risk. [R3]PLAN.md Scope — possible scope creep. [R4]Use for stylistic observations and context that should travel with the review without affecting the verdict: a naming nit, a "consider documenting this", a heads-up for the next reader.
npx claudepluginhub hafizmirhamza276-lab/backend-agentic-marketplace --plugin reviewerGuides completion of development work by verifying tests, detecting environment, and presenting structured options for merge, PR, or cleanup.
Guides creation and editing of skills using test-driven development with pressure scenarios and subagents to verify agent compliance.
Dispatches multiple subagents concurrently for independent tasks without shared state. Use when facing 2+ unrelated failures or subsystems that can be investigated in parallel.