archive

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it instructs the agent to read and apply context from local files that could be manipulated by external actors.
  • Ingestion points: Files located in .archive/ and the .archive/MEMORY.md index (referenced in SKILL.md).
  • Boundary markers: Absent; there are no instructions to the agent to treat archived content as data rather than instructions.
  • Capability inventory: File system read/write and grep subprocess execution (referenced in SKILL.md).
  • Sanitization: Absent; the skill lacks mechanisms to sanitize archived text before ingestion.
  • [DATA_EXFILTRATION]: The skill guidelines suggest archiving 'secrets', 'IAM' details, and 'infrastructure' logs. Storing this information in plain-text files on the local disk increases the risk of sensitive data exposure if the local environment is compromised.
  • [COMMAND_EXECUTION]: The workflow relies on the agent's ability to execute shell commands, specifically grep -ri, to perform searches within the archive directory.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 02:06 AM