The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection.
Ingestion points: Reads and edits files within the project repository as described in SKILL.md and references/implementation-guide.md.
Boundary markers: None identified; repository content is processed without explicit delimiters or instructions to ignore embedded commands.
Capability inventory: Includes Bash, gh, curl, and file Write/Edit permissions as defined in the allowed-tools of SKILL.md.
Sanitization: No sanitization or validation of the ingested repository content is performed before processing.
[DATA_EXFILTRATION]: Performs network requests to api.apollo.io for status checks and integration tests, and to Slack for build alerts. These interactions target well-known services associated with the skill's purpose.
[COMMAND_EXECUTION]: Employs the GitHub CLI (gh) for managing repository secrets and curl for connectivity checks in the workflow templates provided in SKILL.md and references/implementation-guide.md.
[EXTERNAL_DOWNLOADS]: Pulls official and well-known GitHub Actions, such as actions/checkout, actions/setup-node, and slackapi/slack-github-action, from the GitHub Actions marketplace.
[CREDENTIALS_UNSAFE]: Includes a validation script in references/implementation-guide.md that passes an API key in a URL query parameter (?api_key=$APOLLO_API_KEY). This is a best-practice violation as sensitive keys transmitted in URLs may be recorded in server or proxy logs.

apollo-ci-integration