Skills
Navigate
skills/steipete/agent-scripts/agent-transcript/Gen Agent Trust Hub

agent-transcript

Warn

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill accesses sensitive file paths in the user's home directory to retrieve agent logs which may contain session tokens and private data.
  • Evidence: scripts/agent-transcript searches ~/.codex/sessions, ~/.claude/projects, ~/.pi/agent/sessions, and paths within ~/.openclaw.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted session logs which could contain malicious instructions.
  • Ingestion points: The scripts/agent-transcript script reads .jsonl logs from local storage.
  • Boundary markers: Transcripts are delimited by <!-- agent-transcript:start --> and enclosed in quadruple-backtick code blocks.
  • Capability inventory: The skill can write rendered transcripts and HTML previews to the /tmp/ directory via Node.js fs.writeFileSync.
  • Sanitization: Implements a regex-based redact function for secrets and an unsafe pattern check that fails the process if authentication strings or private keys are found.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 16, 2026, 11:26 PM
Security Audit — agent-trust-hub — agent-transcript