The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data from database migrations and Change Data Capture (CDC) events, which represents a vulnerability surface for indirect prompt injection.
Ingestion points: The process_cdc_events method in SKILL.md reads events from Kafka topics, and the executeMigrationWithObservability method processes MongoDB collection data.
Boundary markers: The provided code snippets do not implement explicit delimiters or boundary markers to separate untrusted data from the agent's instructions.
Capability inventory: The skill environment allows for Bash and WebFetch usage, and the code snippets utilize the Python requests library for outbound network communication.
Sanitization: There is no evidence of data sanitization or validation for the incoming event payloads before they are processed or logged.
[EXTERNAL_DOWNLOADS]: The instructions and code snippets reference several standard third-party libraries that would need to be installed in the environment.
Node.js dependencies: mongodb, winston, and prom-client.
Python dependencies: kafka-python, prometheus_client, requests, and numpy.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill includes built-in functionality for transmitting data to external monitoring and alerting services.
The AlertingSystem class is designed to send payloads to external Slack webhooks via requests.post.
The EnterpriseMigrationMonitor class facilitates data transmission to a Grafana API, including the handling of Authorization Bearer tokens.
[COMMAND_EXECUTION]: The skill includes a CI/CD integration section in SKILL.md that defines shell command execution steps for GitHub Actions workflows, such as running monitoring and migration scripts via the command line.

database-migrations-migration-observability