The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to fetch and run the @playwright/cli package from the official npm registry. This is a well-known service used for the verification phase of the workflow.
[COMMAND_EXECUTION]: Shell commands are used to execute Playwright for headed browser walkthroughs, console log capture, and network request monitoring during both the signal-gathering and verification phases.
[DATA_EXFILTRATION]: The skill reads .env.local to identify project-specific identifiers such as SENTRY_ORG and SUPABASE_PROJECT_ID. While this file typically contains sensitive secrets, the instructions explicitly restrict the agent to retrieving configuration names for tool targeting.
[PROMPT_INJECTION]: The skill processes external data from Sentry issue summaries and Supabase logs. This presents an indirect prompt injection surface where malicious content in logs could influence agent behavior.
Ingestion points: Sentry issue searches (sentry:search_issues) and Supabase log retrieval (supabase:get_logs) in SKILL.md.
Boundary markers: No explicit delimiter or instruction-guard markers are defined for the ingested data content.
Capability inventory: SQL execution (execute_sql), database migration (apply_migration), and shell command execution (npx).
Sanitization: Not specified; however, the skill mandates a 'Triage' and 'Plan' phase where the agent must present the intended actions to the user for review before implementation.

iterate-post-launch