Mirrored from filpgf.io — ProPGF Batch 3 (Karma program 1479, application 6a16e6f0e43618775073655d, status: pending). Contact details redacted; canonical application lives on filpgf.io. 1.1 Project Name Aweb Filecoin Agent Receipt Archive 1.2 Project Github https://github.com/manfromnowhere143 1.3 Project Website https://aweb-wine.vercel.app/grants/agent-receipts 1.4 Team Lead/Point of Contact Daniel Wahnich, Founder, Aweb Labs. Preferred contact: Telegram @cogitoergosum143 or email [redacted] 1.5 Category [ "Other", "Core Infrastructure" ] 1.6 Open Source Status Partial 2.1 Project Summary Aweb Filecoin Agent Receipt Archive will define an open-source archive profile for AI-agent execution evidence: scoped authority, tool/provider invocation, status, cost, failure state, recovery state, and redaction boundaries. The Filecoin-specific work packages those receipts into durable, addressable bundles that can be preserved and inspected through Filecoin/IPFS-compatible storage patterns. It supports application builders, storage providers, and network infrastructure teams that need credible records of agent and tool execution beyond volatile app logs or private dashboards. The gap is simple: agents are beginning to perform real work, but their execution evidence is rarely portable, durable, or reviewable after the workflow ends. Filecoin can become the natural evidence archive for public agent runs, benchmark artifacts, audit records, and grant-reportable execution traces. 2.2 Who does this work support? [ "Application Users", "Storage Providers", "Application Builders", "Network Infrastructure" ] 2.3 Total Funding Requested (USD) 180000 2.4 Milestones & Budget [ { "title": "Receipt Archive Profile and Redaction Model", "description": "Define the first Filecoin/IPFS-compatible receipt archive profile and public evidence model.\n\nWork includes the archive manifest, redaction model, bundle structure, metadata fields, privacy boundaries, threat model, sample datasets, and examples for public agent execution receipts. This milestone turns Aweb Agent Receipts from a generic evidence primitive into a storage-aware Filecoin profile: what gets archived, what must be redacted, how bundles are addressed, and how humans can verify integrity without seeing secrets.", "dueDate": "2026-07-31", "fundingRequested": "45000", "completionCriteria": "Completion criteria:\n\n- Public archive profile draft published under Apache-2.0.\n- JSON/schema examples for successful, failed, retried, partial, and human-stopped agent executions.\n- Redaction policy, privacy boundary, and threat model included.\n- At least 20 example receipt bundles prepared for validator and storage work.\n- Clear guidance on what belongs in a public Filecoin/IPFS archive and what must stay private.\n- Reviewer-ready technical note explaining how this supports Filecoin application builders and infrastructure maintainers." }, { "title": "TypeScript CLI, Validator, and Filecoin/IPFS Examples", "description": "Build the reference TypeScript package, CLI validator, fixture generator, and Filecoin/IPFS integration examples.\n\nThe CLI will validate receipt bundles, report missing fields, check redaction requirements, generate archive-ready manifests, and verify retrieved bundles. The examples will show package/address/retrieve flows using Filecoin/IPFS-compatible storage patterns, with fixture data that application builders can run locally without private Aweb credentials. This is the engineering-heavy milestone: SDK, CLI, tests, examples, and integration documentation.", "dueDate": "2026-09-30", "fundingRequested": "75000", "completionCriteria": "Completion criteria:\n\n- TypeScript helper package with typed archive manifest and receipt bundle helpers.\n- CLI commands for validate, inspect, package, retrieve-check, and verify.\n- Filecoin/IPFS-compatible example workflows documented and runnable.\n- Fixture generator creates realistic public agent/tool/API execution bundles.\n- Tests cover success, failure, retry, partial execution, redaction, invalid manifest, missing fields, and corrupted retrieval cases.\n- Public examples run from a clean checkout without private Aweb credentials." }, { "title": "Archive Viewer, Documentation, and Final Report", "description": "Ship a lightweight human viewer, public benchmark-style fixture set, documentation, ecosystem review loop, and final grant report.\n\nThe viewer will let reviewers inspect archive manifests, receipt status, authority, provider/tool references, c …[truncated] 3.1 Impact pathway The contribution is indirect at first and can become direct if downstream builders adopt the archive profile. Output: public Filecoin/IPFS-compatible receipt archive profile, validator, fixture set, examples, and viewer for agent execution evidence. Outcome: application builders, storage providers, and infrastructure teams get a reusable way to preserve AI-agent/tool/API evidence outside one private app, one provider dashboard, or one cloud log. Impact: Filecoin becomes a credible evidence archive for agent runs, benchmark artifacts, grant-reportable traces, public audits, and workflow accountability. That strengthens the network's application-builder surface and makes durable storage relevant to a frontier AI-agent use case rather than only generic file preservation. 3.2 Verification metrics Metric | Data source | How measured | Target Archive profile released | GitHub / public reviewer page | Published spec, schemas, and examples | 1 complete Filecoin/IPFS-compatible profile Validator coverage | Repository tests / CI logs | Passing tests for success, failure, retry, partial, redacted, invalid, and corrupted cases | 50+ validation cases Public fixture set | Repository examples | Receipt/archive bundles across agent, MCP, API, cloud-worker, and message-action scenarios | 30+ bundles Filecoin/IPFS-compatible examples | Public docs / example scripts | Documented package/address/retrieve/verify flow | 3+ end-to-end examples Viewer usability | Public viewer / docs | Human can inspect authority, provider/tool, cost, status, failure, redaction, and recovery fields | Viewer released with sample data Ecosystem feedback | GitHub issues / review notes | Filecoin/IPFS or agent-framework reviewers can test and comment | 5+ external feedback threads targeted 3.3 References No Filecoin-specific beneficiary reference is available yet; this proposal is intentionally scoped as a new public-good infrastructure milestone rather than a maintenance grant for an existing Filecoin dependency. Public technical references for reviewer diligence: - Aweb Agent Receipts public dossier: https://aweb-wine.vercel.app/grants/agent-receipts - Public spec draft: https://aweb-wine.vercel.app/specs/aweb-agent-receipts.md - GitHub profile: https://github.com/manfromnowhere143 If shortlisted, I can provide a technical walkthrough, repo access to relevant private work where safe, or a live call with Filecoin reviewers before final grant agreement. The funded deliverable itself will be public. 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? The grant would fund most of the dedicated Filecoin open-source workstream for six months. The broader Aweb Labs product remains founder-funded and supported by small personal seed capital. 4.3 If this grant is not awarded, what happens? The Filecoin-specific archive profile would slow down and likely remain a local/offchain receipt primitive for longer. The broader Aweb Labs work would continue founder-funded, but the public Filecoin/IPFS evidence archive milestone would be reduced in scope. 4.4 Core Team Daniel Wahnich - founder and solo builder of Aweb Labs. Relevant work: building Aweb as a governed execution layer for AI agents, including model/tool routing, MCP/provider integration, inbox/background-worker operations, public agent receipt specs, and grant-reviewer surfaces. Public materials: https://aweb-wine.vercel.app/grants/agent-receipts and https://github.com/manfromnowhere143. No public LinkedIn profile is available; GitHub and public reviewer pages are the primary technical references. 4.5 Has your team received a ProPGF grant or funding from PLFIF before? [ "No" ] 5.1 Key risks & dependencies Key risks and dependencies: - RFP alignment: Batch 3 emphasizes core infrastructure and RFP responses. This proposal is strongest if Filecoin reviewers agree that durable agent execution evidence is relevant infrastructure for application builders and storage tooling. - Storage integration depth: the milestone should not pretend to be a mature production Filecoin deployment on day one. It will ship archive profiles, validator, fixtures, and integration examples first, then harden based on reviewer feedback. - Privacy: receipt archives must avoid leaking secrets, private prompts, credentials, personal data, or unnecessary payloads. Redaction and threat modeling are first-class deliverables. - Public repo separation: the current Aweb monorepo is private; the grant scope must be extracted into a clean public Apache-2.0 repository/package. - Solo-builder bandwidth: the ask is sized for a serious six-month build with possible external review/support, but the milestone plan remains narrow and inspectable. - Ecosystem validation: downstream value depends on feedback from Filecoin/IPFS builders, agent-framework teams, and potential storage-provider users. Any feedback you have on the application process? The form is clear about public visibility and operational expectations. The sign-in warning is useful because it prevents losing a detailed application after drafting. Anything else you want to share that we didn't ask? Aweb Labs is not asking Filecoin to fund another generic chatbot. The proposal is for durable execution evidence: a public primitive that lets agents leave records that people and systems can actually inspect after the run is over. The broader Aweb thesis is that autonomous systems need governed execution before they can safely do serious work. Filecoin is a strong fit for the archive layer because durable storage matters most when the live workflow is gone and someone needs to prove what happened. The USD 180,000 request is below the Batch 3 soft cap and aligned with the published average grant size, but scoped tightly enough to produce public deliverables within six months. Contributing to Core Infrastructure? New open-source Filecoin/IPFS-aligned infrastructure for archiving and validating agent execution evidence; depends on Filecoin/IPFS-compatible storage services and supports app builders, storage tooling, and public-good reviewers. Objective 1 Indirect Objective 2 Indirect Objective 3 Indirect Open Source Context The funded Filecoin scope will be released as public Apache-2.0 code and documentation: archive profile, redaction rules, TypeScript helpers, CLI validation commands, Filecoin/IPFS-compatible examples, fixtures, viewer, and docs. The current Aweb monorepo is private while the broader product is still being assembled and contains unrelated Aweb Labs work. That private product code is not the requested grant deliverable. If funded, the Filecoin Agent Receipt Archive scope will be separated into a public repository/package under the Aweb Labs/GitHub profile so Filecoin reviewers and builders can inspect, fork, and reuse it.