The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the official @larksuite/cli Node.js package, which is installed using npx if not already present on the system. This is an official utility from a well-known service provider.
[COMMAND_EXECUTION]: The script scripts/setup_app_credentials.py uses the Python subprocess module to configure and interact with the lark-cli tool. It handles the Feishu App Secret by passing it through standard input (stdin) to the CLI command, ensuring the sensitive value is not exposed in the process argument list or system logs.
[CREDENTIALS_UNSAFE]: The skill manages Feishu application credentials (App ID and App Secret). It uses a local YAML configuration file (~/.config/soia-skills/soia-cwork-feishu-cli/config.yml) and provides a dedicated script to initialize the CLI profile. The implementation follows safe handling practices to minimize the risk of credential exposure.
[PROMPT_INJECTION]: The skill processes untrusted data from Feishu documents, creating a surface for indirect prompt injection.
Ingestion points: Document content and metadata fetched via lark-cli commands (e.g., docs +fetch, wiki +node-list).
Boundary markers: The skill does not implement explicit delimiters or specific instructions for the agent to ignore embedded commands within the fetched content.
Capability inventory: The skill uses subprocess.run to execute various lark-cli commands for data retrieval and profile management.
Sanitization: No programmatic sanitization or filtering of the document content is performed before it is presented to the agent.

soia-cwork-feishu-cli