BLAST Execution Validation Skill¶
Motivation¶
BLAST execution validation spans the dashboard submit path, the sibling OpenAPI submit path, Celery queue behavior, Playwright E2E coverage, Azure login state, and Application Insights telemetry. Agents needed a repeatable workflow that starts with safe local checks, escalates to live submits only with explicit scope, and loops through fix/rerun cycles autonomously.
User-Facing Change¶
Added a project-scoped Copilot skill named blast-execution-validation. It can be invoked for BLAST execution test planning, parallel queue validation, OpenAPI submit hardening, App Insights error hunts, and autonomous E2E run/fix/rerun work.
The skill now defines concrete handling for scope, concurrency, and max-hours. Explicit scope: full-azure counts as approval for the guarded lifecycle validation run, while concurrency is reserved for post-lifecycle submit fan-in probes and max-hours is treated as a hard autonomous run budget.
API/IaC Diff Summary¶
- No runtime API or IaC changes.
- Added
.github/skills/blast-execution-validation/SKILL.md. - Added skill references for the scenario matrix, App Insights KQL, and autonomous loop rules.
- Clarified full Azure invocation behavior for
/blast-execution-validation scope: full-azure concurrency=2 max-hours=4.
Validation Evidence¶
- Frontmatter and file placement checked with shell validation.
- The skill references existing validation entry points:
api/tests/test_external_blast_api.py,api/tests/test_blast_submit_route_options.py,api/tests/test_blast_queue.py,api/tests/test_blast_tasks.py,api/tests/test_openapi_rate_limit.py,scripts/dev/e2e-ui.sh, and the Playwright E2E scripts underscripts/e2e/scenarios/.