Skills
Navigate
skills/bahayonghang/my-ai-cli-toolkit/skill-session-review/Gen Agent Trust Hub

skill-session-review

Warn

Audited by Gen Agent Trust Hub on Oct 16, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill accesses sensitive local chat history stores for multiple platforms, specifically ~/.claude/projects/, ~/.grok/sessions/, ~/.codex/sessions/, and ~/.omp/agent/sessions/ through scripts/scan_invocations.py. These files contain historical user conversations which are highly sensitive, although the skill includes a basic redaction mechanism for certain API tokens in scripts/write_session_review.py.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from historical session logs as part of its primary workflow. Instructions or malicious patterns embedded in these logs could potentially influence the agent's behavior when it processes the resulting feedback reports or when it follows the generated handoff prompt defined in references/handoff-prompt.md.
  • [COMMAND_EXECUTION]: The skill executes local Python scripts (scripts/scan_invocations.py and scripts/write_session_review.py) and Git commands (git rev-parse, git check-ignore) as defined in the allowed-tools section of SKILL.md.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 16, 2026, 03:26 AM