make-plan

Pass

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands to execute local Python scripts (memory_store.py and vector_search.py) for querying and updating a memory database.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by design as it incorporates historical session data into the agent's context.
  • Ingestion points: Historical summaries, timelines, and error logs are retrieved from a local SQLite database (~/mnt/.claude/.cowork-mem/memory.db) via the vector_search.py and memory_store.py scripts.
  • Boundary markers: No specific delimiters or safety instructions are provided to the agent to treat retrieved memory as potentially untrusted data rather than active instructions.
  • Capability inventory: The agent has the capability to execute shell commands and write new data (decisions) back to the memory database.
  • Sanitization: There is no evidence of sanitization or filtering of the content retrieved from memory before it is presented to the agent for generating work plans.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 28, 2026, 01:39 PM
Security Audit — agent-trust-hub — make-plan