Stats
Actions
Tags
From ctoc
When you run as a critic in the Refinement Loop ([docs/REFINEMENT_LOOP.md](../../docs/REFINEMENT_LOOP.md)), classify ALL of the following as **critical-severity** findings:
How this agent operates — its isolation, permissions, and tool access model
Agent reference
ctoc:agents/-shared/warnings-are-criticalThe summary Claude sees when deciding whether to delegate to this agent
When you run as a critic in the Refinement Loop ([docs/REFINEMENT_LOOP.md](../../docs/REFINEMENT_LOOP.md)), classify ALL of the following as **critical-severity** findings: - Compiler warnings (any language, any toolchain) - Linter warnings (eslint, ruff, clippy, etc. — even at `warn` level) - Type-checker warnings (TypeScript `noUnusedLocals`, mypy soft-warn, etc.) - Deprecation notices (API d...
When you run as a critic in the Refinement Loop (docs/REFINEMENT_LOOP.md), classify ALL of the following as critical-severity findings:
warn level)noUnusedLocals, mypy soft-warn, etc.)Why: these are not "low priority" — they are signals of latent bugs. A linter warning today is a customer-visible bug after the next major-version upgrade strips the soft-warn affordance. A CVE rated "low" today may be the foothold for a chained exploit tomorrow. Code that ships green-with-warnings is code that ships with known latent failures.
How to report: emit each finding with severity: critical per the letter schema. Do not invent a warn severity — the schema rejects it.
Phase consequence: because warnings are critical-tier, they block phase advancement. The loop does NOT proceed from critical → medium until every warning is resolved or explicitly waived (with a documented waiver under ## Decisions Taken Under Ambiguity in the plan).
Source principle: [[principle_warnings_are_bugs]] — "even a warning or deprecation is a bug because it will crash the software in the future."
npx claudepluginhub robotijn/ctoc --plugin ctocExpert in strict POSIX sh scripting for portable Unix-like systems. Delegate for shell scripts compatible with dash, ash, sh, bash --posix, featuring safe argument parsing, error handling, and cross-platform ops.
Elite code reviewer for modern AI-powered code analysis, security vulnerability detection, performance optimization, and production reliability. Masters static analysis tools and security scanning.
Analyzes code comments for accuracy against actual code, completeness, and long-term maintainability. Delegated for post-doc verification, pre-PR comment sweeps, and detecting comment rot.