fix(evidence): bound project filters to FTS candidates

[solomonneas]

Summary

• Keep the materialized FTS candidate set as the outer query loop for filtered searches.
• Evaluate project metadata only for bounded candidates instead of scanning every archived item first.
• Cover exact and substring project filters and assert candidate-first query planning.

Measured result

The project-filtered MCP query used during issue #501 work fell from more than 224 seconds to 0.09 seconds on the same archive.

Verification

• brigade work verify run --target . --command "go -C engines/evidence-ledger vet ./..." --capture brigade-work
• brigade work verify run --target . --command "go -C engines/evidence-ledger test ./..." --capture brigade-work
• brigade work verify run --target . --command "env GOFLAGS=-modcacherw engines/evidence-ledger/scripts/smoke_mcp.sh" --capture brigade-work

[coderabbitai]

Important

Review skipped

Auto reviews are limited based on label configuration.

🏷️ Required labels (at least one) (1)

• coderabbit-review

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository: escoffier-labs/coderabbit/.coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 213f876e-67a5-45ee-9418-bba746a54717

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/evidence-project-filter-search-index

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}