discover-my-skills
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from the user's chat history, transcripts, and files to generate new agent instructions (skills). This process lacks boundary markers or sanitization to prevent the agent from inadvertently treating malicious instructions embedded within the history as valid rules for the new skill.
- Ingestion points:
SKILL.mdspecifies reading the current conversation, supplied transcripts, files, and using tools to search or read other chat sessions. - Boundary markers: Absent; the skill does not instruct the agent to delimit or ignore embedded instructions found in the processed history.
- Capability inventory: The skill possesses the capability to write new files to the filesystem, as defined in
SKILL.md("create the actual skill files") andreferences/creation-and-evidence.md("Use the current environment's file-creation capabilities"). - Sanitization: Absent; there is no requirement to validate or filter the content extracted from history before interpolating it into the new skill's instructions.
Audit Metadata