The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted conversational data from call transcripts during the scoring process.
Ingestion points: Call recordings and transcripts are pulled from Nooks and Clari as defined in the source inventory.
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious commands embedded within the processed transcripts.
Capability inventory: The skill has the ability to send Slack messages and create Gmail drafts, which could be misused if an injection is successful.
Sanitization: No content filtering or sanitization of external transcript text is specified.
[COMMAND_EXECUTION]: The skill references using the bash date command to calculate the weekly report window. This is a standard and safe implementation for time-based logic.
[EXTERNAL_DOWNLOADS]: The skill retrieves data from Nooks and Clari MCP servers. These are described as recognized internal integrations necessary for the skill's stated business objectives.

sdr-call-coaching