The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local bash script (scripts/buffer.sh) which in turn calls standard system binaries curl and jq. This is the primary mechanism for interacting with the Buffer API.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources (Buffer social media posts and ideas) that may contain content provided by third parties.
Ingestion points: Data is ingested through the list-posts, get-post, and channels functions in scripts/buffer.sh which query the Buffer GraphQL API.
Boundary markers: The output is returned as structured JSON, which provides implicit boundaries, though no specific 'ignore instructions' markers are added to the text fields.
Capability inventory: The skill has network access via curl to api.buffer.com and processes data using jq.
Sanitization: The skill uses jq's argument handling (--arg, --argjson) to safely construct JSON payloads, preventing command injection during API request formation. It does not sanitize the text content retrieved from the API before presenting it to the agent.

buffer