test: make Proof Pack runnable (harness pipeline tests)#33
Conversation
|
J’ai poussé des commits supplémentaires pour faire passer les checks qui échouaient :
Ces commits sont sur la branche |
|
Nouveau commit pour fiabiliser les 2 checks restants :
Commit HEAD: |
|
Nouveau commit pour corriger
Commit HEAD: |
Removed broken workflow badge links that were causing link-check failures: - secret-scan.yml badge - link-check.yml badge
Fixed broken MCP Server Gmail link causing link-check failure. Changed from: - https://github.com/anthropics/mcp-servers (404) To: - https://github.com/modelcontextprotocol/servers (correct repo)
|
J’ai committé le fix Commit HEAD: |
Fixed typo in MCP Server Gmail link. Changed from: - https://github.com/modelcontextprotocolserv/ers (404) To: - https://github.com/modelcontextprotocol/servers (correct)
|
Fix appliqué : section Commit HEAD: |
Resolves merge conflict with main by: - Restoring CI badges with absolute URLs (fixes link-check) - Adding missing opening ``` for ASCII art block 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Adds end-to-end pytest for benchmark harness pipeline - Makes harness scripts configurable via env vars - Adds proof-pack.yml workflow - Adds .lychee.toml config - Adds benchmarks/compare.sh script - Adds scripts/proof.sh and scripts/run-tests.sh 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
✅ Merged manuallyLes changements de cette PR ont été intégrés manuellement sur Fichiers intégrés:
🤖 Generated with Claude Code |
Summary
BENCH_RESULTS_JSONLBENCH_METRICS_DIRBENCH_SUMMARY_JSONBENCH_BADGES_DIRWhy
This turns the "Proof Pack" from a purely narrative requirement into a runnable, CI-verifiable pipeline, even before application code exists.