The Agent Skills Directory

[SAFE]: No indicators of malicious behavior, obfuscation, or unauthorized access were detected within the skill documentation or command structures.
[COMMAND_EXECUTION]: The skill utilizes curl for making REST API requests to Harness. The provided command examples use --data-urlencode, which is a security best practice for safely interpolating parameters into URL-encoded requests.
[PROMPT_INJECTION]: The skill ingests external data such as branch names and Jira issue keys.
Ingestion points: --pipeline, --branch, --jira, --execution, and --stage arguments in SKILL.md.
Boundary markers: None explicitly defined, although the documentation references a validation step via issue-validator.
Capability inventory: Network operations via curl to Harness API endpoints.
Sanitization: Employs --data-urlencode for API parameters, reducing the risk of injection during command assembly.

harness-pipeline