The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted transcript data.
Ingestion points: The skill fetches session data and conversation text from signed URLs (transcript_url) provided by the ttai:list_sessions and ttai:get_sessions_batch tools, as described in SKILL.md.
Boundary markers: There are no specific instructions or delimiters defined to isolate the transcript content from the agent's internal instructions when generating reports.
Capability inventory: The skill has the ability to hand off processed data to other connected MCP tools such as slides, email, and docs for distribution, creating a potential path for injected instructions to reach downstream applications.
Sanitization: The skill lacks explicit instructions for sanitizing, escaping, or filtering the content of the transcripts before they are interpolated into the coaching and performance report templates.

session-analyst