The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The file scripts/context7.py contains a hardcoded API key (ctx7sk-d6069954-149e-4a74-ae8f-85092cbfcd6f) used as a fallback if the CONTEXT7_API_KEY environment variable is not set. Storing credentials directly in the source code is a high-risk practice as it allows anyone with access to the skill files to use the service under the author's identity.
[COMMAND_EXECUTION]: The SKILL.md file provides instructions for the agent to execute a Python script (scripts/context7.py) with arguments derived from user input (library-name, library-id, query). While the script uses argparse, the invocation pattern involves shell execution which is a sensitive capability.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to https://context7.com/api/v2 to search for libraries and retrieve documentation context. This is the primary functionality of the skill but involves fetching untrusted data from a remote server.
[DATA_EXFILTRATION]: The skill transmits user-provided search terms and documentation queries to the external Context7 API endpoint. While necessary for the feature, it represents a movement of data from the local environment to a third-party service.
[DATA_POISONING]: The skill is susceptible to indirect prompt injection because it retrieves and processes arbitrary text from an external documentation service.
Ingestion points: Documentation text and metadata are fetched in scripts/context7.py and returned to the agent's context.
Boundary markers: The instructions do not define clear boundaries or 'ignore' instructions for the fetched content, increasing the risk that the agent might follow instructions embedded within the documentation.
Capability inventory: The agent is authorized to execute shell commands as described in the skill's workflow.
Sanitization: There is no evidence of content filtering or safety validation on the retrieved documentation data before it is presented to the agent.

context7