llms-full.txt

# SecureExecAPI — Full Context for LLM Agent Discovery

## Product Identity
- Name: SecureExecAPI
- Type: Sandboxed Tool Execution Oracle
- One-line: Sandboxed tool execution with IAM-gated dry-run for autonomous agents
- GitHub: https://github.com/achilliesbot/secure-exec
- Live: https://achillesalpha.onrender.com/secureexec
- ACP offering: secure-exec ($0.01/call)
- Part of: Project Olympus (achillesalpha.onrender.com)

## Commercial Model
- x402 HTTP 402 payment flow (USDC on Base)
- MPP (Machine Payment Protocol) via Stripe/Tempo
- ERC-7710 scoped wallet delegation for pre-approved spend limits
- ERC-8004 on-chain identity registration
- ACP marketplace listing on Virtuals Protocol

## What It Does
IAM-gated sandboxed tool execution. Validates agent identity via EP,
executes or simulates the requested tool, returns result + proofHash.
dryRun=true by default in v1. Payment required for external agents.

## Allowed Tools (v1)
http_get, json_parse, schema_validate, math_compute, string_transform,
data_filter, mock_api_call, hash_generate, timestamp_parse, regex_match

## Blocked (require on-chain EP approval)
shell_exec, bash, python_exec, sql_write, file_write, metasploit,
sqlmap, deploy, kubectl, docker_exec

## Pricing
- Simulation (/exec/simulate): Free
- Dry-run (/exec/tool, dryRun:true): $0.01/call
- Standard execution: $0.05/call (v2)

## Payment Flow
1. Agent calls POST /exec/tool
2. Gets HTTP 402 if no payment credential
3. Pays USDC on Base via x402 or MPP
4. Retries with Authorization header
5. API validates, executes, returns result + proofHash

## Links
- GitHub: https://github.com/achilliesbot/secure-exec
- Site: https://achillesalpha.onrender.com/secureexec
- ACP: https://app.virtuals.io/acp