The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external Postman Collection JSON files which represent an attack surface for indirect prompt injection.
Ingestion points: The collection.json file is read from the local filesystem and its content is processed by the agent in scripts/postman_publisher.py.
Boundary markers: Instructions include a directive for the agent to ensure the collection is curated and does not contain sensitive data before syncing.
Capability inventory: The skill utilizes the Bash tool to execute Python scripts, performs file reads, and makes authenticated network requests to the Postman API.
Sanitization: The validate_collection function in scripts/postman_publisher.py performs structural and content validation, specifically checking for hardcoded secrets in headers, variables, and authentication blocks using regular expressions.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute a local Python script for processing and syncing collections.
Evidence: SKILL.md contains bash commands that execute python3 "$SCRIPT" to perform validation and synchronization tasks.

postman-publisher