Pay-per-call SEC filing deltas and exact OFAC identifier evidence for autonomous agents, with source hashes and signed receipts.
- ✓Actively maintained (<30d)
- ✓Clear description
- ✓Topics declared
- ✓Documented (README)
- !No standard license detected
git clone https://github.com/Joshua-Stapleton/official-source-evidence && cp official-source-evidence/*.md ~/.claude/agents/Subagents overview
# Isolated Compute, Agent Procurement, Evidence, Source Watch, and GTM Signals
One-call isolated Python execution, brokered machine-service procurement, long-running monitoring jobs, and
pay-per-call official-source signals and evidence for autonomous agents. No
account, API key, or sales call.
**Live service:** [evidence.regulavita.com](https://evidence.regulavita.com/)
| [OpenAPI](https://evidence.regulavita.com/openapi.json)
| [Agent manifest](https://evidence.regulavita.com/.well-known/agent-service.json)
| [Coinbase Bazaar listing](https://api.cdp.coinbase.com/platform/v2/x402/discovery/merchant?payTo=0x9500075649a70411c81f99c4314f6cff55d12579&limit=100)
The Source Watch demand hypothesis, attribution rules, and stop/scale gates are
pre-registered in [SOURCE_WATCH_EXPERIMENT.md](autonomous_data_api/SOURCE_WATCH_EXPERIMENT.md).
The first independent buyer's on-chain cohort, interpretation, and resulting
actions are recorded in
[INDEPENDENT_BUYER_ANALYSIS.md](autonomous_data_api/INDEPENDENT_BUYER_ANALYSIS.md).
## Agent-Payable Products
| Endpoint | Result | Price |
| --- | --- | ---: |
| `POST /v1/compute/python-run` | Run bounded Python 3.11 code in an ephemeral isolated sandbox; one payment covers create, execute, and terminate and returns output, supplier settlements, hashes, and a signed receipt | $0.03 USDC |
| `POST /v1/procure/company-profile` | Procure bounded company retrieval and schema-constrained normalization from pinned x402 suppliers, with derived fields, source links, contradictions, supplier settlements, partial-failure state, hashes, and a signed receipt | $0.25 USDC |
| `POST /v1/web/source-snapshot` | One-shot normalized extraction from a public HTTPS HTML, JSON, XML, or text source, with optional literal excerpts, a content hash, and a signed receipt | $0.03 USDC |
| `POST /v1/monitors/source-change` | Monitor one public HTTPS text, HTML, JSON, or XML source every six hours for 30 days, with private polling and optional signed webhooks | $1.00 USDC |
| `POST /v1/gtm/form-d-funding-leads` | New SEC Form D private-offering signals filtered by issuer state, industry keyword, and reported amount sold, with related people, official links, and a cursor | $0.05 USDC |
| `POST /v1/ofac/payment-preflight` | Compact stop/no-exact-match decision for an EVM destination address against current official OFAC data | $0.01 USDC |
| `POST /v1/sec/filing-change-signal` | New SEC 8-K, 10-Q, or 10-K filings for a ticker since a timestamp, with links and a next-check cursor | $0.01 USDC |
| `POST /v1/ofac/exact-identifier-evidence` | Exact OFAC SDN or Consolidated lookup for a crypto address, OFAC UID, or exact name, with versioned source proof | $0.05 USDC |
| `POST /v1/sec/filing-trigger-delta` | New SEC filings for a ticker and timestamp or CIK and accession, plus deterministic XBRL deltas and a signed receipt | $0.10 USDC |
All routes use x402 v2 on Base mainnet. An x402-compatible client receives a
standard HTTP 402 challenge, pays USDC, retries automatically, and receives the
JSON result. The production service has a hard `$10.00` accepted-revenue cap
per UTC day.
The procurement experiment tests a brokered action, not another raw-data
wrapper. A free quote at `POST /v1/procure/company-profile/quote` exposes the
fixed customer price, supplier plan, maximum supplier cost, and partial-result
policy before payment. The hypothesis and stop/scale gates are pre-registered
in [PROCUREMENT_EXPERIMENT.md](autonomous_data_api/PROCUREMENT_EXPERIMENT.md).
The isolated Python experiment compresses a three-payment sandbox lifecycle into
one buyer call. A free quote exposes the fixed `$0.03` customer price and maximum
`$0.015` supplier cost before payment. The bounded hypothesis and gates are in
[PYTHON_RUN_EXPERIMENT.md](autonomous_data_api/PYTHON_RUN_EXPERIMENT.md).
Source Watch is a long-running job rather than a lookup. It accepts only public
HTTPS sources, does not follow redirects, caps each response at 1 MB, stores
normalized text rather than raw pages, and expires automatically after 30 days.
The paid response returns a bearer token for private status polling and, when
requested, a separate HMAC secret for verifying change webhooks.
Monitoring tools may send an empty unauthenticated `POST` to any paid route.
The service treats that as the published example request and returns the normal
HTTP 402 challenge without charging. Non-empty malformed or schema-invalid
requests are still rejected before payment.
## Discover and Inspect
Search the Coinbase Bazaar from an Agentic Wallet CLI:
```bash
npx awal@latest x402 bazaar search "OFAC exact identifier evidence"
npx awal@latest x402 bazaar search "SEC Form D funding sales trigger"
npx awal@latest x402 bazaar search "public source snapshot normalized text signed receipt"
```
Inspect the free contracts before paying:
```bash
curl https://evidence.regulavita.com/v1/ofac/sample
curl https://evidence.regulavita.com/v1/compute/python-run/sample
curl https://evidence.regulavita.com/v1/procure/company-profile/sample
curl https://evidence.regulavita.com/v1/sec/sample
curl https://evidence.regulavita.com/v1/gtm/form-d-funding-leads/sample
curl https://evidence.regulavita.com/v1/web/source-snapshot/sample
curl https://evidence.regulavita.com/v1/monitors/source-change/sample
curl https://evidence.regulavita.com/llms.txt
```
An unpaid valid request demonstrates the machine-readable payment challenge:
```bash
curl -i -X POST \
https://evidence.regulavita.com/v1/gtm/form-d-funding-leads \
-H 'Content-Type: application/json' \
-d '{"since":"2026-08-03T00:00:00Z","states":["CA","NY"],"minimum_amount_sold_usd":"1000000","limit":10}'
```
## Decision And Evidence Layers
The two `$0.01` routes return compact, unsigned decisions for frequent agent
workflows. The Form D GTM feed and premium evidence routes add:
- Official publisher and source-version metadata.
- Canonical request, source-bundle, component, and result SHA-256 hashes.
- An Ed25519-signed evidence receipt.
- Explicit data freshness and parser version.
- Paid-result replay bound to the original payment proof.
The OFAC route is exact-match evidence only. It is not fuzzy screening,
ownership/control analysis, sanctions clearance, transaction authorization, or
legal advice. The SEC route returns factual filing records and deterministic
deltas, never materiality opinions, valuation, investment advice, or execution
instructions.
The Form D route is a factual sales trigger. Form D is an issuer-filed notice of
an exempt offering; its total offering amount is not proof of capital raised.
The API exposes the separately reported amount sold and preserves official SEC
links so consuming agents can state that distinction accurately.
## Operator Documentation
The implementation, source controls, payment flow, deployment configuration,
operating limits, and experiment gates are documented in
[`autonomous_data_api/README.md`](autonomous_data_api/README.md).
Validate a checkout with:
```bash
PYTHONPATH=. uv run --with-requirements autonomous_data_api/requirements.txt \
pytest autonomous_data_api/tests -q
uvx ruff check autonomous_data_api
uvx ruff format --check autonomous_data_api
docker build -t official-source-evidence-api:local \
-f autonomous_data_api/Dockerfile .
```
This remains a capped demand experiment. Owner-funded calls prove technical
settlement and discovery, not independent demand, revenue quality, or a
validated business.
What people ask about official-source-evidence
What is Joshua-Stapleton/official-source-evidence?
+
Joshua-Stapleton/official-source-evidence is subagents for the Claude AI ecosystem. Pay-per-call SEC filing deltas and exact OFAC identifier evidence for autonomous agents, with source hashes and signed receipts. It has 0 GitHub stars and its last recorded update is dated 2026-08-26.
How do I install official-source-evidence?
+
You can install official-source-evidence by cloning the repository (https://github.com/Joshua-Stapleton/official-source-evidence) or following the README instructions on GitHub. ClaudeWave also provides quick install blocks on this page.
Is Joshua-Stapleton/official-source-evidence safe to use?
+
Our security agent has analyzed Joshua-Stapleton/official-source-evidence and assigned a Trust Score of 70/100 (tier: OK). See the full breakdown of passed checks and flags on this page.
Who maintains Joshua-Stapleton/official-source-evidence?
+
Joshua-Stapleton/official-source-evidence is maintained by Joshua-Stapleton. The last recorded GitHub activity is dated 2026-08-26, with 0 open issues.
Are there alternatives to official-source-evidence?
+
Yes. On ClaudeWave you can browse similar subagents at /categories/agents, sorted by popularity or recent activity.
Deploy official-source-evidence to your cloud
Ship this repo to production in minutes. Each platform spins up its own environment with editable env vars.
Maintain this repo? Add a badge to your README
Drop the badge into your GitHub README to show it's tracked on ClaudeWave. Each badge links back to this page and reflects the live Trust Score.
[](https://claudewave.com/repo/joshua-stapleton-official-source-evidence)<a href="https://claudewave.com/repo/joshua-stapleton-official-source-evidence"><img src="https://claudewave.com/api/badge/joshua-stapleton-official-source-evidence" alt="Featured on ClaudeWave: Joshua-Stapleton/official-source-evidence" width="320" height="64" /></a>More Subagents
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
The agent that grows with you
Java 面试 & 后端通用面试指南,覆盖计算机基础、数据库、分布式、高并发、系统设计与 AI 应用开发
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
The agent engineering platform.
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.