The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install well-known API testing tools from the standard NPM registry.
Evidence: Recommends installing newman and newman-reporter-htmlextra using npm install -g.
[COMMAND_EXECUTION]: The skill generates shell scripts and Jenkinsfile pipelines intended to run the newman command for API testing.
Evidence: Provides reusable bash scripts and Groovy pipeline scripts that execute the newman run command with various parameters for collection execution and reporting.
[CREDENTIALS_SAFE]: The generated scripts use secure patterns for handling sensitive information such as API keys and environment files.
Evidence: Recommends using Jenkins credentials store (credentials('id')) and environment variable placeholders ({{POSTMAN_API_KEY}}) instead of hardcoding secrets.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for processing external data files which could theoretically contain malicious instructions, though it currently only generates the execution logic.
Ingestion points: Processes external files such as collection.json, environment.json, and test-data.csv in SKILL.md.
Boundary markers: Absent from the generated command templates.
Capability inventory: Executes subprocesses via the newman CLI tool.
Sanitization: None explicitly defined for file paths or URLs, relying on the user to provide valid testing assets.

postman-newman-automation