The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the mcp__scheduled-tasks__create_scheduled_task tool to establish automated, recurring persistence for the competitive intelligence scan.
[COMMAND_EXECUTION]: Local state and persistent configuration (watchlist and baselines) are maintained by writing files to the local directory ~/.claude/competitor-watch/.
[DATA_EXFILTRATION]: Strategic reports and competitive insights are transmitted to external Slack channels using the mcp__a8f5bb61-0837-408d-a165-744ad0d8d236__slack_send_message tool.
[PROMPT_INJECTION]: The skill is exposed to indirect prompt injection via the processing of untrusted external content from ad transcripts.
Ingestion points: Creative transcripts are fetched from external ad sources using the mcp__motion__get_creative_transcript tool.
Boundary markers: The instructions do not provide explicit delimiters or instructions to treat external transcript content as data rather than instructions.
Capability inventory: The agent has the ability to write to the file system, send messages via Slack, and create scheduled tasks.
Sanitization: No filtering or validation of the fetched transcript text is performed before the content is processed for strategic theme extraction.

competitor-watch