The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to retrieve and process external data from the Fathom API, creating a potential surface for indirect prompt injection attacks where malicious instructions hidden in transcripts could influence agent behavior.
Ingestion points: Meeting transcripts and AI-generated summaries are fetched from the api.fathom.ai endpoints as described in the Step 2 and Step 3 bash instructions in SKILL.md.
Boundary markers: The instructions do not define any delimiters or provide warnings to the agent to treat the retrieved text as untrusted data or to ignore embedded instructions.
Capability inventory: The skill allows the agent to use Read, Write, Edit, and Bash tools, which could be leveraged if an attacker embeds malicious instructions within a recorded meeting transcript.
Sanitization: The provided commands use jq to structure the JSON output, but no sanitization, filtering, or escaping is performed on the actual text content of the meeting segments or summaries.

fathom-hello-world