# Caelara Labs Agent Safety Suite

> Zero-scrape guardrails for AI agents: input firewall, output validator, PII redaction, prompt-injection scanning, token compression, JSON repair, schema enforcement, grounding diff, citation extraction and context chunking. Pay per call: x402 USDC on Base (see https://agent.caelaralabs.com/.well-known/x402), a direct X-Caelara-Key, or RapidAPI.

- **Domain**: `agent.caelaralabs.com`
- **Provider**: Caelara Labs (https://caelaralabs.com)
- **A2A protocol version**: 1.0
- **Kind**: paid_api
- **Live-responds (last probe)**: True
- **Signed card**: False
- **Streaming**: False
- **Quality score**: 100%
- **Payment protocol**: x402 · network `eip155:8453`
- **Wallet**: `0x085910da18a735bac73b8ddb58a06a99a510ebc6`

## Skills
- **Scan untrusted input: injection risk + PII redaction + sanitized text + allow/review/block verdict** — Scan untrusted input: injection risk + PII redaction + sanitized text + allow/review/block verdict. POST /v1/agent-input-firewall. Parameters: text* (body): Unt
- **Validate LLM output: repair JSON, enforce schema, check grounding against a source, extract citations** — Validate LLM output: repair JSON, enforce schema, check grounding against a source, extract citations. POST /v1/agent-output-validator. Parameters: output* (bod
- **Redact emails, phones, SSNs, cards (Luhn), IPs, API keys and JWTs** — Redact emails, phones, SSNs, cards (Luhn), IPs, API keys and JWTs. POST /v1/pii-redactor. Parameters: text* (body): Text to redact. 0.005 USDC per call (atomic 
- **Score text for prompt-injection patterns** — Score text for prompt-injection patterns. POST /v1/prompt-injection-scanner. Parameters: text* (body): Text to scan. 0.005 USDC per call (atomic tier).
- **Compress text to a target ratio of its original length** — Compress text to a target ratio of its original length. POST /v1/token-compressor. Parameters: text* (body): Text to compress; ratio (body): Target ratio 0.1-0.
- **Repair malformed JSON from LLM output** — Repair malformed JSON from LLM output. POST /v1/json-auto-repair. Parameters: text* (body): Malformed JSON text. 0.005 USDC per call (atomic tier).
- **Validate and coerce data to a JSON Schema** — Validate and coerce data to a JSON Schema. POST /v1/schema-enforcer. Parameters: data* (body): Data to enforce; schema* (body): JSON Schema subset. 0.005 USDC p
- **Flag answer sentences unsupported by the source text** — Flag answer sentences unsupported by the source text. POST /v1/hallucination-diff. Parameters: source* (body): Source / ground-truth text; answer* (body): Model
- **Extract URLs, DOIs, arXiv ids and author-year citations** — Extract URLs, DOIs, arXiv ids and author-year citations. POST /v1/citation-extractor. Parameters: text* (body): Text containing citations. 0.005 USDC per call (
- **Split text into token-bounded overlapping chunks** — Split text into token-bounded overlapping chunks. POST /v1/context-chunker. Parameters: text* (body): Text to chunk; max_tokens (body): Max tokens per chunk (de

## URLs
- Agent card: https://agent.caelaralabs.com/.well-known/agent-card.json
- Page (HTML): https://agenstry.com/agents/agent.caelaralabs.com
- Documentation: https://agent.caelaralabs.com/openapi.json
