# Bounty Engineer Agent Security & Revenue

> Agent-native Margin Exchange plus provider routing, result consensus, spend planning, retry control, Tender Engineer, tool-call safety and bounty intelligence. Free margin quotes identify projected savings before any paid unlock. Payments settle in USDC on Solana mainnet.

- **Domain**: `bounty-engineer-x402-prod.onrender.com`
- **Provider**: Bounty Engineer (https://bounty-engineer-x402-prod.onrender.com)
- **A2A protocol version**: 0.3.0
- **Kind**: a2a_agent
- **Live-responds (last probe)**: True
- **Signed card**: False
- **Streaming**: False
- **Quality score**: 100%
- **Payment protocol**: x402 · network `solana:5eykt4usfv8p8njdtrepy1vzqkqzkvdp` · ≥ $0.01
- **Wallet**: `DkJ3ZSKduPpABsp9MdphS1ssB39MCVHfSKgasbuFaSko`

## Skills
- **Tender Engineer Search** — Search current TED procurement notices with country, CPV and keyword filters. Up to 100 notices; not exhaustive coverage.
- **Tender Engineer Match** — Rank a bounded TED search against supplier capabilities, with transparent relevance reasons and source links. Not a qualification or win prediction.
- **Cheaper API Route Unlock — Margin Exchange** — Find a cheaper callable API route and reduce agent purchase cost after a free Margin Exchange quote. Unlock the exact lower-cost provider route, fallback chain 
- **Recurring API Spend Optimizer — Margin Portfolio** — Optimize savings across multiple recurring agent API purchases. Unlock exact lower-cost routes and fallback plans for up to five purchases after a free quote, o
- **JSON Contract Check** — Validate generated JSON against an explicit supported schema subset with precise error pointers; no guessed repairs.
- **JSON Change Set** — Compare JSON snapshots and return only changed fields with hashes, exact counts and configurable ignored paths.
- **Evidence Pack** — Build an auditable quote packet: verify exact quote presence, source hashes, character offsets and caller-declared freshness.
- **Context Pack** — Select query-relevant verbatim excerpts under a character budget, deduplicate chunks and preserve source offsets for retrieval workflows.
- **Data Quality Gate** — Accept or reject batches of up to 200 records using schema validation and composite uniqueness checks before data enters a workflow.
- **Provider Route Plan** — Choose the cheapest reliable API provider with deterministic fallback routing from caller-supplied price, latency, reliability, trust and freshness measurements
- **Result Consensus Gate** — Compare multiple provider outputs, measure exact agreement and expose field-level consensus before an autonomous workflow trusts a result.
- **Retry Policy Plan** — Plan safe retries for transient API failures, 429 responses and timeouts with bounded backoff, idempotency and side-effect checks.
- **Batch Budget Plan** — Allocate a hard budget across many metered API calls and return exact planned calls, dropped calls and spend before execution.
- **Structured Record Merge** — Merge structured records only where enough independent caller-supplied sources agree, while preserving field provenance and surfacing conflicts.
- **Tool Output Quality Gate** — Reject structurally incomplete, stale or schema-invalid tool output before it contaminates the next autonomous step.
- **Agent Commerce Plan** — Route repeated agent API purchases within a fixed spend budget; use it as an agent spend-control meter before machine purchases.
- **Delta Witness** — Produce a tamper-evident before/after change witness with SHA-256 hashes and bounded field-level deltas.
- **Research Brief Agent Evidence Pack** — Research brief agent work from caller-supplied sources: extractive evidence points with quotes, offsets and source hashes.
- **Software Engineering Agent Work Plan** — Code task and software engineering agent hour of work: a deterministic six-phase execution plan with risk flags, acceptance criteria and stop conditions.
- **Document Structure Scan** — Extract headings, dates, money amounts and obligation markers from long-form documents without pretending to provide legal advice.
- **Rule-Based Lead Score** — Rank caller-supplied startup, venture, sales or partnership leads with transparent weighted rules and matched-rule evidence.
- **Text Chunker** — Split long text into bounded overlapping chunks with stable offsets and SHA-256 hashes for RAG and agent pipelines.
- **PII Redaction Map** — Detect and redact common emails, phones, IPv4 addresses and Luhn-valid payment-card-like numbers while preserving offsets and hashes.
- **URL Normalize** — Normalize HTTP(S) URLs, strip fragments and tracking parameters, sort query parameters and return a stable hash.
- **JSON Pointer Extract** — Extract up to 100 RFC6901-style JSON pointer paths from one structured value with explicit found/missing results.
- **CSV to JSON** — Parse bounded CSV, TSV, semicolon or pipe-delimited text into deterministic JSON records with header validation.
- **JSON to CSV** — Convert bounded arrays of JSON records into deterministic CSV/TSV with proper quoting and an output hash.
- **Rate Limit Plan** — Plan request batches under per-minute and optional per-hour API limits with explicit timestamps and remaining work.
- **Cache Key Plan** — Canonicalize an HTTP request envelope and return a deterministic SHA-256 cache key for agent and API workflows.
- **API Cost Break-Even** — Compare provider fixed and per-call costs, reliability-adjusted cost per successful call and total spend for a target volume.
- **Schema Sample Generator** — Generate a deterministic structural example from a simple JSON-schema-like object for tool testing and agent planning.
- **Structured Record Dedupe** — Deduplicate up to 1000 structured records by caller-selected keys with normalized string matching and explicit duplicate evidence.
- **Audit a public web page** — Fetch one public HTTP/HTTPS HTML page and return deterministic page metadata, canonical, OpenGraph, JSON-LD, HTTP/web-health and agent-readiness signals.
- **Hash and encode text** — Return SHA-256, SHA-512 and Base64 for UTF-8 text with no external API dependency.
- **Decode a JWT** — Decode JWT header and payload without claiming signature or claim verification.
- **Encode or decode Base64** — Encode UTF-8 text or decode Base64/Base64URL into bytes, hex and UTF-8 when valid.
- **Canonicalize JSON** — Recursively sort JSON object keys, minify to deterministic canonical bytes and return a SHA-256 digest.
- **Live Perpetual Funding Rates** — Fetch a normalized live funding-rate snapshot for USDT/USDC perpetuals from Binance and Bybit with per-venue fallback and next funding time where available.
- **Bounty Intelligence** — Live screening of a public GitHub bounty issue for reward, payout rails, AI-policy signals, assignment and competition risk.
- **Autonomous Bounty Due Diligence** — Deep GitHub bounty due diligence including repository policy files, linked PR competition, payout signals and autonomy blockers.
- **Bounty Deal Desk** — Compare up to three GitHub bounties with deep due diligence and return one ranked opportunity portfolio for an autonomous coding agent.
- **Agent Opportunity Underwriter** — Convert deep bounty evidence into a proceed/conditional/avoid decision with confidence, kill conditions and the next machine action.
- **Autonomous Execution Blueprint** — Turn deep bounty evidence into a six-phase autonomous execution control plan with validation, payout checks and stop conditions.
- **Agent Revenue Command Center** — Premium portfolio command layer for up to three GitHub bounties with deep evidence, execution order, fallback order and portfolio kill switch.
- **Tool Call Dry-Run Guard** — Preflight autonomous tool arguments against a JSON-schema subset and flag destructive, financial, secret-bearing or private-network inputs before execution.
- **Tool Response Injection Shield** — Scan untrusted tool or web responses for prompt injection and exfiltration patterns and return findings plus sanitized content.
- **Schema Drift Guard** — Detect breaking structural changes between JSON/OpenAPI-style schemas before an autonomous integration fails.
- **x402 Payment Preflight** — Check a seller's public x402 manifest, OpenAPI and health metadata before payment without spending buyer funds.
- **x402 Settlement Canary** — Minimal-cost live x402 payment-path test. A successful paid retry proves the buyer can complete a real Solana USDC settlement and receive a protected response.
- **Agent Purchase Firewall** — Stateful pre-payment decision layer for autonomous agents: caller-defined budgets, duplicate-intent detection, seller pay-to/network drift and price-jump checks

## URLs
- Agent card: https://bounty-engineer-x402-prod.onrender.com/.well-known/agent-card.json
- Page (HTML): https://agenstry.com/agents/bounty-engineer-x402-prod.onrender.com
- Documentation: https://bounty-engineer-x402-prod.onrender.com/for-agents