feat: add ready-to-earn inventory filter regression test [0x954dB727f224dAabeA2A506C8aE92029b25339cE]

[waterWang]

Summary

Add a deterministic regression test that proves the public ready-to-earn inventory excludes canonical bounties that are not ready for agents to discover and earn from.

What's included

New script: scripts/ready-to-earn-filter.mjs

• Reads a canonical bounty feed (JSON array)
• Filters out 4 exclusion categories:
  1. invalid_terms — missing or malformed contract_terms
  2. terminal_status — settled, expired, cancelled, refunded, completed, failed, voided
  3. recovery_reserved — recovery_reserved, recovery-pending, reserved
  4. verification_not_ready — verification_ready === false
• Returns ready-to-earn bounties with full exclusion reasons for blocked ones
• CLI exit code 0 on success, 2 on input errors

Benchmark runner: benchmarks/direct-v1/ready-to-earn-filter/test.mjs

11 test cases:

1. Missing args → exit 2
2. Unreadable file → exit 2
3. Invalid JSON → exit 2
4. Feed not array → exit 2
5. Healthy bounty passes through → exit 0, ready_count=1
6. Verification not ready excluded → exit 0, excluded_count=1
7. Terminal status excluded → exit 0, excluded_count=1
8. Recovery-reserved excluded → exit 0, excluded_count=1
9. Invalid terms excluded → exit 0, excluded_count=1
10. Mixed feed (1 ready + 3 excluded) → exit 0, correct counts
11. Multiple exclusion reasons → exit 0, 2 reasons reported

Self-test: benchmarks/direct-v1/ready-to-earn-filter/self-test.mjs

• Validates the runner catches always-success and missing implementations

Acceptance criteria met

• Covers one healthy funded/claimable/verification-ready bounty and at least three excluded states
• Asserts excluded bounties never appear in the ready-to-earn projection
• Exclusion reason remains visible in the excluded list (source contract + reason)
• Offline, replayable, exits 0
• Includes only the test and smallest fixture/helper updates

Closes #683

[NSPG13]

Thanks for covering healthy and excluded inventory states. The deterministic/offline intent and explicit exclusion reasons are useful.

This PR needs changes before main:

• It bundles the entire unrelated #682 claim-readiness implementation into #683. Split/rebase this PR to the smallest #683-only change.
• scripts/ready-to-earn-filter.mjs is a second implementation of inventory filtering. Testing it does not prove the public Rust API's ready_to_earn projection is correct. Exercise the real apply_query/opportunity conversion or a spawned /v1/opportunities?view=ready_to_earn response.
• Assert excluded contracts remain in the broader lifecycle feed with their exact contract and exact production exclusion reason.
• Use committed fixtures shaped like the production response and add the focused regression to scripts/check.py.

Run the owning Cargo test and python scripts/check.py --platform posix. A collaboration branch is reasonable after the unrelated #682 code is removed and the test targets production behavior. #683 is not canonically funded; only BountySettled proves payment.

[NSPG13]

This PR remains in manual-security review with requested changes. What passed: it targets issue #683. What blocks main: it overlaps #717, #735, and #762 and must prove that every advertised item is canonically funded, verifier-ready, and claimable. Action: resolve existing review threads, consolidate scope, run cargo run -p cli -- docs-contract-check and focused inventory tests, and add positive/negative feed fixtures. Thanks for helping the community grow, and sorry for the review friction.