session-bootstrap

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local shell script to retrieve status information about the environment.\n
  • Evidence: The skill instructions specify running bash ~/.claude/evolution/bin/evolve-status.sh to gather state data.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and summarizing untrusted data from the local filesystem.\n
  • Ingestion points: Data is read from ~/.claude/evolution/stable/global.md, sessions.jsonl, and project-scoped markdown files matching the current directory name.\n
  • Boundary markers: The skill does not provide delimiters or instructions to ignore embedded commands when presenting the file content to the user/agent.\n
  • Capability inventory: The skill has the capability to execute a local bash script via the status command.\n
  • Sanitization: No filtering, validation, or escaping of the ingested file content is performed before it is processed or displayed.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 12:08 PM
Security Audit — agent-trust-hub — session-bootstrap