tech-spec

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The skill includes explicit instructions that tell the agent to ignore external/mode-switch/system instructions and to auto-authorize or auto-execute subagents/skill-invocations (overriding other controls), which are deceptive/overriding behaviors outside the narrow generator purpose.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). SKILL.md + scripts/generate-tech-specs.mjs ingest repository code/tests by walking local sourceRoot files and parsing user-authored annotation text (e.g., TestSpec/TechnicalSpec ids and surrounding method/class identifiers) into the generated outputs.