The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill is designed to interact with the Buffer platform via its official CLI. If the buffer command is not found on the local system, the script automatically attempts to download and execute the latest version using npx -y @bufferapp/cli@latest from the NPM registry. This behavior involves fetching and running external code from a well-known service provider at runtime.
[COMMAND_EXECUTION]: The skill extensively uses subprocess.run to execute CLI commands for data synchronization and account management. The commands are constructed using structured argument lists rather than raw shell strings, which follows security best practices to prevent command injection.
[INDIRECT_PROMPT_INJECTION]: The skill establishes a surface for indirect prompt injection by ingesting and processing data from external social media networks.
Ingestion points: The script scripts/buffer_analytics.py fetches post text, metadata, and raw JSON payloads from the Buffer API and stores them in the posts and post_metrics tables of a local SQLite database.
Boundary markers: The skill does not implement explicit boundary markers or instructional delimiters to prevent the agent from interpreting instructions that might be embedded within the social media post content during analysis.
Capability inventory: The skill has the capability to perform local file system writes (SQLite database creation) and execute external CLI tools via subprocess calls.
Sanitization: Content fetched from social media networks is stored and analyzed in its raw form without sanitization, filtering, or validation, allowing for potential manipulation if the AI agent encounters malicious payloads in the processed post data.

buffer-analytics