pi-history-ingest

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external text files containing user conversations, tool results, and bash outputs from prior sessions, introducing an indirect prompt injection surface.
  • Ingestion points: Processes JSONL history files from ~/.pi/agent/sessions/ or the path specified by PI_HISTORY_PATH in SKILL.md (Step 1 and Step 2).
  • Boundary markers: There are no explicit structural boundary tags or escaping instructions defined for isolating raw text fields when loaded into the model context.
  • Capability inventory: The skill writes files to the Obsidian vault path ($OBSIDIAN_VAULT_PATH) and invokes external CLI tools (find and qmd).
  • Sanitization: Incorporates defensive mechanisms requiring the removal of API keys, tokens, credentials, and raw long tool payloads prior to ingestion.
  • [COMMAND_EXECUTION]: The skill directs the agent to execute shell utilities locally to find files and update indexes.
  • Evidence: Utilizes find to index history files and triggers ${QMD_CLI:-qmd} update and ${QMD_CLI:-qmd} embed to refresh the local QMD search index.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 03:07 AM
Security Audit — agent-trust-hub — pi-history-ingest