LIFECYCLE HOOKS

ARTIFICIAL INTELLIGENCE ASSISTANT PROMPTS Prompt 1 helps you find gaps in hook coverage. [START OF PROMPT] Review this harness's hook implementations (pre-execution, post-execution, validation, approval, continuation) for coverage gaps: [PASTE LAB hooks.py / HOOK REGISTRATIONS HERE]. For each hook type missing or incompletely implemented, specify exactly what check it should add and why its absence is risky. Output format: a table with the columns Hook Type, Present (Y or N), Missing Check, and Risk. [END OF PROMPT] Prompt 2 helps you design and test a continuation hook. [START OF PROMPT] Design a continuation hook that detects premature surrender in agent output (phrases like "I cannot do this" or "this is too complex") and re-injects the goal with unused strategies: agent/task = [AGENT DESCRIPTION, e.g. Resume Review Agent matching resumes to postings]. Then test the hook's detection logic against this transcript and tell me whether it would have fired: [PASTE TRANSCRIPT HERE]. Output format: hook design with pseudocode, then a pass or fail test result with justification. Include a cap on the number of continuations and explain why it is needed. [END OF PROMPT]