Skills
Navigate
skills/first-fluke/oh-my-agent/oma-recap/Gen Agent Trust Hub

oma-recap

Pass

Audited by Gen Agent Trust Hub on Oct 5, 2026

Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill accesses sensitive local files containing AI conversation history.\n
  • Evidence: It reads from ~/.claude/history.jsonl and uses the oma recap tool to extract conversation data for summarization.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted conversation data which could contain malicious instructions.\n
  • Ingestion points: AI conversation logs extracted via oma recap --json and the file ~/.claude/history.jsonl (SKILL.md).\n
  • Boundary markers: There are no explicit boundary markers or instructions to ignore embedded prompts when the agent synthesizes the history into a recap.\n
  • Capability inventory: The skill utilizes oma recap and jq to read history and writes the generated Markdown recaps to .agents/results/recap/ (SKILL.md).\n
  • Sanitization: The jq processing logic includes minimal sanitization by replacing newlines with spaces and truncating text to 150 characters (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 5, 2026, 04:14 PM