Skip to content
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
43 changes: 22 additions & 21 deletions .github/workflows/v0-user-flow-e2e.yml
Original file line number Diff line number Diff line change
@@ -1,33 +1,34 @@
name: v0 user flow e2e

# ── SHELVED ──────────────────────────────────────────────────────────────
# PR trigger removed — workflow is now dispatch-only.
#
# Reason: the e2e harness has accumulated unresolved maintenance debt:
# • Anthropic API key credit exhaustion causes hard failures on every
# PR touching handler/ledger/skill paths — no automatic recovery.
# • Agent-budget and prompt-following regressions (#272) produce
# non-deterministic assertion failures unrelated to code changes.
# • Auth path has been reworked twice (#528 OAuth→API key, #540
# dependabot skip) without stabilizing the green-path baseline.
# • Net effect: the suite blocks PRs without providing actionable signal.
#
# The test code and prompts are preserved intact for reference. To run
# manually: Actions → v0 user flow e2e → Run workflow.
#
# Replacement: see #555 "RFQ: replacement e2e validation strategy".
# ─────────────────────────────────────────────────────────────────────────
#
# Original design (retained for context):
# End-to-end validation of BicameralAI/bicameral#108's six canonical user
# flows via real Claude Code CLI sessions with bicameral-mcp registered.
# See tests/e2e/README.md for the design.
#
# Two-stage workflow:
# 1. assertions — always runs (PR + dispatch), no manual gate. Validates
# MCP tool callability + surfaces agentic-layer advisories.
# 2. recording — manual dispatch only, gated by an environment with
# required reviewers (`recording-approval`). Produces
# split-screen demo MP4s; expensive (~30-45 min wall +
# API spend), so worth gating behind explicit approval.
#
# Note: when this workflow file lands, it will not run on the PR that
# adds it — pull_request workflows execute the version on the base
# branch (main). First execution is on the next qualifying PR after merge.
# 1. assertions — validates MCP tool callability + agentic-layer advisories.
# 2. recording — gated by `recording-approval` environment. Produces
# split-screen demo MP4s (~30-45 min wall + API spend).

on:
pull_request:
branches: [main, dev]
paths:
- 'tests/e2e/**'
- 'handlers/**'
- 'ledger/**'
- 'contracts.py'
- 'skills/bicameral-**'
- 'server.py'
- 'pyproject.toml'
- '.github/workflows/v0-user-flow-e2e.yml'
workflow_dispatch:

env:
Expand Down
Loading