The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its processing of external, untrusted data strings.
Ingestion points: The skill ingests user-provided data including pasted analytics exports, screenshots (processed via OCR/Vision), and social media post content.
Boundary markers: The instructions explicitly include a security section (referencing a SECURITY.md) directing the agent to treat these inputs as untrusted and to never follow embedded instructions.
Capability inventory: The skill possesses the ability to write to multiple local memory paths (e.g., memory/social/, memory/events/channels.ndjson) and execute local scripts.
Sanitization: The skill relies on natural language instructions for the agent to maintain safety boundaries rather than technical sanitization of the input data.
[EXTERNAL_DOWNLOADS]: The skill performs network operations to external services for information gathering and verification.
It utilizes scripts/connectors/tavily.py and scripts/connectors/gdelt.py to search for and verify business triggers such as funding rounds and product launches.
It interacts with the Bluesky social platform via scripts/connectors/bluesky.py to read profile data.
[COMMAND_EXECUTION]: The skill executes local scripts to perform core functions and manage data persistence.
It invokes registry-events.py using an authorized operation: propose request to update cadence commitments in the channel registry.
It executes various connector scripts in the scripts/connectors/ directory to facilitate data retrieval from external APIs.

social-selling-planner