pi-history-ingest
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to facilitate the data ingestion and indexing process.
- Uses
findto locate session JSONL files within the local Pi history directory. - Executes
obsidian-wiki memory syncto update the vault's manifest and activity logs. - Runs
qmd updateandqmd embedto maintain the search index for the Obsidian vault. - [INDIRECT_PROMPT_INJECTION]: The skill ingests session history data which may contain instructions that could influence the agent's behavior.
- Ingestion points: Reads structured JSONL session files from
~/.pi/agent/sessions/(SKILL.md). - Boundary markers: The instructions do not specify the use of delimiters or markers to distinguish between the session data and the agent's instructions, increasing the risk of the agent following commands embedded in the logs.
- Capability inventory: The skill has the ability to write to the filesystem (Obsidian vault) and execute several shell commands (
find,obsidian-wiki,qmd). - Sanitization: While the skill instructs the agent to remove credentials and redact sensitive text, it does not explicitly sanitize for adversarial instructions embedded within the conversation turns.
Audit Metadata