The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains a Python script designed to be executed in a Django management shell ('managerie') to modify production database records (cancelling stuck AsyncTask records). While this represents a high-privilege operation, it is consistent with the skill's primary purpose of remediating platform incidents and targets the vendor's own infrastructure.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the interpolation of untrusted data into SQL queries and Python scripts.
Ingestion points: User-provided workspace_id and contract_id are interpolated into BigQuery SQL statements and Django ORM queries in SKILL.md.
Boundary markers: No delimiters or instructions are used to separate user-provided identifiers from the executable code/query logic.
Capability inventory: The skill facilitates SQL execution via BigQuery and arbitrary Python execution via the Django shell.
Sanitization: There is no explicit validation or escaping of the input identifiers before they are inserted into the command strings.
[EXTERNAL_DOWNLOADS]: The skill references several external resources, including SpotDraft admin panels, Sentry logs, and BigQuery projects (spotdraft-prod, spotdraft-qa). These references are neutral, point to the vendor's official infrastructure, and are necessary for the documented troubleshooting workflow.

metadata-extraction-task-stuck