prior-work-retrieval

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONPERSISTENCEDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/prior_work.py script executes system commands including ripgrep (rg) for searching the filesystem and git for checking repository status. It also supports a "command" carrier mode that executes programs defined in a local JSON manifest file.
  • [PERSISTENCE]: The skill includes an installation script scripts/prior-work-retrieval.sh that registers hooks in the AI agent's configuration files (~/.claude/settings.json and ~/.codex/hooks.json). These hooks allow the skill to intercept user prompts and tool usage to enforce its retrieval workflow across sessions.
  • [DYNAMIC_EXECUTION]: The scripts/prior_work.py script dynamically imports its companion hook script scripts/prior_work_hook.py using Python's importlib.util to access shared classification logic and patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 03:06 AM
Security Audit — agent-trust-hub — prior-work-retrieval