ai-memory-durable-pages

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary purpose is to store project knowledge and rules that persist across sessions, creating an attack surface where instructions stored in one session could influence or override agent behavior in future sessions.
  • Ingestion points: As described in SKILL.md, the agent ingests content from durable wiki pages, repository-level instruction files, and the .ai-memory.toml configuration file.
  • Boundary markers: The instructions in SKILL.md do not specify the use of delimiters or 'ignore embedded instructions' warnings when retrieving or processing stored memory content.
  • Capability inventory: SKILL.md identifies capabilities for writing and deleting wiki pages via memory_write_page and memory_delete_page tools, as well as the ability to update canonical agent instruction files in the local repository.
  • Sanitization: There are no instructions for sanitizing, escaping, or validating the content before it is stored in the durable memory or interpolated into future prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 08:42 PM
Security Audit — agent-trust-hub — ai-memory-durable-pages