Mirrored from filpgf.io — ProPGF Batch 3 (Karma program 1479, application 6a289cb056982ce6278b2ec9, status: pending). Contact details redacted; canonical application lives on filpgf.io. 1.1 Project Name libp2pxfilecoin: Verifiable Multi-Agent Coordination (py-libp2p + Filecoin) 1.2 Project Github https://github.com/acul71/libp2pxfilecoin 1.3 Project Website https://github.com/acul71/libp2pxfilecoin 1.4 Team Lead/Point of Contact Luca Pisani — py-libp2p co-maintainer | email: [redacted]| Telegram: @acul71 | GitHub: https://github.com/acul71 1.5 Category [ "Other" ] 1.6 Open Source Status Fully Open Source 2.1 Project Summary RFP: Filecoin ProPGF Batch 3 — Call for Builders (June 2026). Focus Area 3: "AI infrastructure products on Filecoin" — verifiable agent audit trails, tamper-evident multi-agent coordination (https://www.filecoin.io/blog/ProPGF-Batch-3-Call-for-Builders). libp2pxfilecoin: verifiable multi-agent coordination on py-libp2p for adversarial settings — signed negotiation FSM, LLM verification pipeline, adversarial tests, trace export. Filecoin/IPLD stores tamper-evident audit transcripts (content-addressed, verifiable retrieval). libp2p handles live coordination; Filecoin handles durable audit provenance. Grant window: 2026-09-01 to 2027-02-28 (6 months). September = grant setup + CI bootstrap; M1–M4 checkpoints through February. Solo part-time. $18,000 total ($17,500 labor + $500 M3 calibnet FIL). Releases v0.2.0 (M2) and v1.0.0 (M4). Aligns with RFS-1, RFS-2, RFS-5, RFS-7. Objective 3 (Indirect): official KPI = flagship client paid onchain adoption; our leading indicators = reference tooling + 2 builder teams on store_audit (see §3.1). Lead: Luca Pisani (acul71). 2.2 Who does this work support? [ "Application Builders", "Application Users", "Network Infrastructure" ] 2.3 Total Funding Requested (USD) 18000 2.4 Milestones & Budget [ { "title": "Grant setup + CI bootstrap + ARENA baseline", "description": "September 2026 (grant month 1): onboarding, milestone plan in repo, CI bootstrap (pytest adversarial suite, ruff), ARENA traceability checklist draft (threat model, 3 scenarios, pass/fail criteria), stable protocol IDs documented (/arena/negotiation/1.0.0), two-agent local demo recorded (arena-negotiation CLI), docs/GRANT_SCHEDULE.md published. No PyPI release yet — prepares M2 v0.2.", "dueDate": "2026-09-30", "fundingRequested": "4500", "completionCriteria": "- CI workflow runs on main (pytest adversarial suite + ruff)\n- ARENA checklist draft in docs/ARENA.md (threat model + 3 scenarios + pass/fail)\n- Baseline two-agent negotiation demo recorded and linked in docs/\n- docs/GRANT_SCHEDULE.md with 6-month plan 2026-09-01 → 2027-02-28\n- Protocol IDs for negotiation, execution, audit documented in docs/\n- Test harness or docker-compose stub committed for reviewer reproduction" }, { "title": "v0.2 release + RFS-5 fee-gated reference", "description": "October–November: Fee-gated messaging design mapped to Filecoin payment primitive (design doc + self-contained local mock — milestone sign-off does not depend on py-libp2p #1286 merge). Optional integration sketch with #1286 if merged before M2. LLM verification hardening — extended invariant tests + CI regression. Benchmark note: negotiation latency, verification failure rate, abort codes. Release v0.2.0 on PyPI with CHANGELOG.", "dueDate": "2026-11-30", "fundingRequested": "4500", "completionCriteria": "- Local fee-gate mock in docs/rfs5/ (independent of #1286):\n - Input: negotiation envelope + payment-token stub\n - Pass: message forwarded only when token valid; ABORT_FEE_GATE logged when invalid\n - Demo script exits 0; expected log lines in docs/rfs5/demo.md\n- RFS-5 design doc published in docs/rfs5/\n- Git tag v0.2.0 + CHANGELOG.md; v0.2.0 on PyPI\n- LLM verification regression suite green in CI\n- Benchmark note published in docs/ (latency, rejection rate, abort codes)\n- CI green on main at sign-off" }, { "title": "Filecoin audit persistence + store/fetch API", "description": "December–January: Audit export → IPLD pipeline (signed transcript bundles). Minimal store_audit / fetch_audit Python API. One paid audit manifest pin on Filecoin calibration network ($500 FIL direct cost within this $4,500 milestone). Recorded demo: negotiate → execute → export → pin → fetch audit. Mesh Devices integration pilot attestation (metric 1). Calibnet chosen for low-cost reproducible proof; full deal automation is post-grant non-goal.", "dueDate": "2027-01-31", "fundingRequested": "4500", "completionCriteria": "- docs/filecoin/audit-ipld.md + example CAR; export script produces manifest CID\n- store_audit / fetch_audit API documented with usage example\n- Calibnet pin tx URL …[truncated] 3.1 Impact pathway Official 2026 Network Objective 3 (form label): "Scale Paid Onchain Flagship Client Adoption." Official KPI (per 2026 Filecoin Network Strategy): increase high-credibility organizations running paid, onchain production workloads — AI pipelines, critical datasets, large-scale data platforms — not experimental pilots. This grant contributes Indirectly: we ship reference tooling that lowers onboarding friction for agent builders who may become flagship clients post-grant. PROGRAM KPI (what Filecoin measures network-wide): More flagship institutions with sustained paid onchain storage/retrieval for production workloads. OUR GRANT-LEVEL LEADING INDICATORS (what we measure in 6 months): A) Reference implementation: store_audit/fetch_audit API + calibnet paid pin + v1.0.0 operator guide B) Builder adoption signal: 2 teams running paid audit CAR workflow (leading indicator — not the program KPI itself) C) Outreach log: 5 teams contacted, 2 calls, 1 public integration PR/discussion (M4) LADDER (how B/C plausibly feed the program KPI): Agent builders adopt store_audit for tamper-evident negotiation logs → paid calibnet/mainnet retention becomes routine → successful builder deployments become reference cases for flagship conversion (credibility flywheel per 2026 strategy). We do not claim flagship client count in this $18K scope. HANDOFF CHAIN (outputs → network action): 1. M3: store_audit/fetch_audit API + calibnet paid pin (tx URL, manifest CID, retrieval proof) 2. Mesh Devices (github.com/acul71/meshdevices): store_audit after sessions → paid audit CAR → fetch_audit on verification (leading indicator team 1/2 — under our control) 3. M4: v1.0.0 operator guide + docs/quickstart-external-runner.md (pip/docker one-command calibnet pin/fetch) + outreach 4. Leading indicator team 2/2: external builder walkthrough (Manusheel-facilitated intro if available) OR any non-team GitHub confirmation they ran external-runner path Why Filecoin: Content-addressed IPLD audit bundles + paid storage + verifiable retrieval. RFP Area 3. RFS-1, RFS-5, RFS-2. Milestone outputs: M1 CI+ARENA; M2 v0.2+RFS-5; M3 IPLD+pin+Mesh Devices; M4 v1.0.0+external-runner+outreach. 3.2 Verification metrics Program KPI (Objective 3): flagship client paid onchain adoption (network-level; out of scope to hit directly in 6 months at $18K). Grant-level leading indicators we will measure: 1) Mesh Devices integration (leading indicator team 1/2) — M3 — REQUIRED Data source: github.com/acul71/meshdevices issues How measured: Public issue — M3 runbook (negotiate → export → store_audit → pin → fetch_audit); manifest CID + calibnet tx URL Target: Issue closed by 2027-01-31 2) Second adoption signal (leading indicator team 2/2) — M4 — REQUIRED (fallback if no external team converts) Primary: non-team builder completes operator guide OR runs docs/quickstart-external-runner.md; public GitHub comment/discussion with environment + result Fallback (fully under our control): external-runner quickstart shipped (pip install + one-command calibnet store/pin/fetch) AND outreach log shows ≥5 teams contacted + ≥2 calls; adoption signal = any non-team GitHub confirmation OR recorded demo with named attendee Target: By 2027-02-28 3) M4 outreach plan Data source: docs/outreach/m4-log.md How measured: 5 teams contacted; ≥2 integration calls held; ≥1 public integration PR or discussion post Target: Log published by 2027-02-28 4) Calibnet audit pin (M3) Data source: Calibration network explorer + audit-demo.md How measured: Tx URL; manifest CID; fetch_audit output; FIL cost; payer wallet Target: All in docs/filecoin/audit-demo.md by 2027-01-31 Note: Paid onchain pin (demo wallet), not free pin — reviewer-verifiable on-chain 5) IPLD audit pipeline Data source: docs/filecoin/audit-ipld.md + repo artifacts How measured: Export script produces manifest CID from sample negotiation session Target: By 2027-01-31 (M3) 6) store_audit / fetch_audit API Data source: API docs + demo runbook How measured: Round-trip store → pin → fetch documented and reproducible Target: By 2027-01-31 (M3) 7) RFS-5 fee-gate reference (M2) Data source: docs/rfs5/demo.md + local mock script How measured: Mock pass/fail per M2 completion criteria (token valid → forward; invalid → ABORT_FEE_GATE) Target: By 2026-11-30 8) Adversarial test suite Data source: GitHub CI How measured: 3 adversarial scenarios pass on main at each milestone sign-off Target: 100% pass by 2026-09-30 (M1) and maintained through M4 9) Product releases Data source: PyPI + git tags How measured: v0.2.0 (M2), v1.0.0 (M4) Target: 2 releases 3.3 References 1. Manusheel Gupta — Filecoin / FHE ecosystem mentor; advised on ProPGF Batch 3 and RFS alignment. Invited reviewer on M4 erc-8004-mapping PR; can facilitate external builder intro for M4 outreach (metric 2). GitHub: https://github.com/seetadev 2. Paul Robinson (pacrob) — py-libp2p maintainer and reviewer; can vouch for our maintainer contributions, interop work, and Filecoin networking context. GitHub: https://github.com/pacrob 3. Johanna Moran — libp2p/Filecoin ecosystem adviser (pro bono); can speak to our sustained maintainer role across py-libp2p and py-multiformats. Discord: j0hmo Integration pilot: Mesh Devices (github.com/acul71/meshdevices) — same builder commits to M3 audit runbook attestation (metric 1). 4.1 Monthly Operating Burn [ "< $10K (basic solo operation or part-time team)" ] 4.2 What % of total team monthly burn depends on this grant? 75 4.3 If this grant is not awarded, what happens? libp2pxfilecoin continues as unpaid part-time research code. MVP stays usable locally; v0.2/v1.0 releases, Filecoin audit API, RFS-5 reference, and operator guide slip 6–12 months. 4.4 Core Team Luca Pisani (acul71) — sole builder, part-time over 2026-09-01 to 2027-02-28. Effort model: sequential milestones M1→M4; September explicitly allocated to grant setup (M1). No paid staff. Budget ($18,000 total = $17,500 labor + $500 direct costs): M1 $4,500 (labor): CI + adversarial suite; ARENA checklist + baseline demo; grant schedule docs M2 $4,500 (labor): RFS-5 design + fee-gate mock; LLM verification hardening; v0.2.0 release M3 $4,500 ($4,000 labor + $500 direct): IPLD pipeline + store_audit/fetch_audit API; calibnet pin demo; $500 calibnet FIL + tooling M4 $4,500 (labor): operator guide; ERC-8004 mapping + PR review; v1.0.0 release + outreach support Labor: ~516 hours over 26 weeks (~20 h/week solo part-time, ~86 h/month). $17,500 labor across M1–M4; $500 direct in M3 (calibnet FIL, included in $18,000 total). GitHub Actions free tier. 75% of project-focused effort depends on this grant (§4.2). Experience: py-libp2p co-maintainer (~627 commits); built libp2pxfilecoin MVP (negotiation FSM, LLM verification pipeline, adversarial tests, existing BLE gateway module, trace export). Adviser: Manusheel Gupta (Filecoin / FHE). 4.5 Has your team received a ProPGF grant or funding from PLFIF before? [ "No" ] 5.1 Key risks & dependencies - LLM unpredictability — multi-stage verification pipeline; ABORT on failure - Filecoin integration complexity — M3 scoped to audit manifest pin, not full deal automation - py-libp2p request/response (#1286) — optional M2 integration sketch only; M2 sign-off uses self-contained local mock - Solo builder — public repo, CI, named references mitigate bus factor - Scope — non-goals: full FVM agent economy, mainnet deal automation, commercial SLA Any feedback you have on the application process? AI-generated review WAR! Anything else you want to share that we didn't ask? Track: ProPGF Batch 3 General Track — RFP Response (Category: Other). RFP: "AI infrastructure products on Filecoin" — Focus Area 3, verifiable agent audit trails / multi-agent coordination. Timeline: Start 2026-09-01; M1 checkpoint 2026-09-30 (September work explicit); end 2027-02-28. Objective 3 (Indirect — primary): KPI = 2 teams on store_audit/Filecoin audit CARs by grant end. Handoff: Mesh Devices (M3) + external team via M4 outreach → paid calibnet retention → path to flagship-class agent workloads. Separate from py-libp2p maintainer grant (Core Infrastructure, 202606/). Complementary to Gautam devgrants #2078 (application-layer DX); this is infrastructure-grade negotiation + Filecoin audit product. Non-goals: new BLE development, full FVM economy, mainnet deal automation, commercial SLA. Repo: github.com/acul71/libp2pxfilecoin Objective 1 N/A Objective 2 N/A Objective 3 Indirect Open Source Context Fully open source (MIT/Apache-2.0 planned)