The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The file references/services.md contains multiple hardcoded cleartext credentials in a service map, specifically ADMIN_PASSWORD values for the 'Alexander AI Support Dashboard' and 'EcDash' services.
[COMMAND_EXECUTION]: The script scripts/railway_api.py uses the subprocess.run function to execute shell commands (curl) to interact with the Railway GraphQL API.
[DATA_EXFILTRATION]: The skill identifies and accesses a sensitive authentication token at a specific system path (/root/.secrets/railway_token) in both SKILL.md and scripts/railway_api.py to authenticate API requests.
[PROMPT_INJECTION]: The skill implements a workflow to fetch deployment logs via the deploymentLogs query. This ingested data from an external source lacks boundary markers or sanitization, creating an indirect prompt injection surface where malicious log content could influence agent behavior.
Ingestion points: deploymentLogs GraphQL query in SKILL.md
Boundary markers: Absent
Capability inventory: Shell command execution via subprocess.run in scripts/railway_api.py; environment variable modification via variableCollectionUpsert mutation
Sanitization: Absent; log messages are processed as raw text

railway-ops