The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands to perform API requests and suggests the use of a package runner for initialization.
Evidence: Execution of curl -sS -H "Authorization: Bearer <key>" "<baseUrl>/api/v1/<route>?account=<account>" for data retrieval.
Evidence: Instructions for the user to execute npx @catalyst-cloud/catalyst-skills join.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download and execution of a remote Node.js package during the setup process.
Evidence: Reference to the @catalyst-cloud/catalyst-skills package via npx.
[DATA_EXFILTRATION]: The skill reads sensitive credential data from the local file system and transmits it to an external URL. While this is functional for the skill's purpose, it establishes a pattern of sensitive data access and transmission.
Evidence: Reads ~/.config/catalyst-cloud/customer.json containing baseUrl and key.
Evidence: Sends the extracted key as a Bearer token in the Authorization header to the configured baseUrl.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API endpoints that could contain adversarial instructions intended to influence the agent's behavior.
Ingestion points: Data is retrieved from /api/v1/issues, /api/v1/search, /api/v1/changes, and other tenant routes (SKILL.md).
Boundary markers: The skill instructs the agent to summarize and cite sources but does not employ specific structural delimiters to isolate untrusted data from instructions.
Capability inventory: The agent has the ability to execute shell commands (curl) and read local configuration files.
Sanitization: There is no evidence of sanitization or validation performed on the content returned from the API before it is processed by the agent.

concierge