The Agent Skills Directory

[COMMAND_EXECUTION]: The skill scripts execute shell commands and database queries to perform administrative tasks such as creating access keys and updating user passwords.
Evidence: scripts/agent.py uses subprocess.run to call the AWS (aws iam) and Azure (az ad) CLI tools. scripts/process.py uses database connectors (psycopg2, mysql.connector) to execute ALTER USER SQL statements.
[INDIRECT_PROMPT_INJECTION]: The skill provides an automated health verification feature that ingests a list of URLs and performs network requests, creating a surface for indirect influence or SSRF if the configuration is controlled by an untrusted source.
Ingestion points: The verify_service_health functions in both scripts/agent.py and scripts/process.py accept a list of endpoint configurations containing target URLs.
Boundary markers: No boundary markers or delimiters are used to wrap the ingested URL data to prevent command or prompt injection during processing.
Capability inventory: The skill possesses network access capabilities (requests.get), command execution capabilities (subprocess.run), and database access capabilities.
Sanitization: The logic lacks validation for the URL scheme, hostname, or destination before performing health check requests, which could lead to Server-Side Request Forgery (SSRF) if the input list is poisoned.
[METADATA_POISONING]: The skill exhibits a discrepancy in its identity metadata, which can mislead users or automated systems regarding its origin.
Evidence: The YAML frontmatter in SKILL.md lists the author as 'mahipal', whereas the LICENSE file attributes copyright to 'mukul975'.

performing-service-account-credential-rotation