generated: '2026-09-19'
method: searched
source: >-
  The provider publishes its own Agent Skill: SKILL.md (agentskills.io frontmatter, version 2.6.0, MIT) and an
  AGENTS.md in github.com/teodorofodocrispin-cmyk/TrustBoost-PII-Sanitizer, distributed on ClawHub as
  trustboost-pii-sanitizer. Both are saved verbatim below (method: searched). One further skill was generated
  by API Evangelist for the read-side flow the provider's skill does not cover, grounded in operationIds
  verified against openapi/trustboost-dev-openapi.json.
skills:
- file: trustboost-dev-pii-sanitizer-SKILL.md
  name: trustboost-pii-sanitizer
  method: searched
  source: https://raw.githubusercontent.com/teodorofodocrispin-cmyk/TrustBoost-PII-Sanitizer/main/SKILL.md
  http_status: 200
  fetched: '2026-09-19'
  version: 2.6.0
  license: MIT
  distribution: https://clawhub.ai/teodorofodocrispin-cmyk/skills/trustboost-pii-sanitizer
  description: Provider-published skill — sanitize text with TrustBoost before it reaches an LLM, with the payment model, data-handling notice, access modes, response envelopes and language examples the provider wrote.
  operations: 2
  operation_ids: [sanitize_pii, sanitize_preview]
  note: >-
    Verbatim. It carries the provider's own transparency notices (raw text goes to OpenAI GPT-4o-mini; "learning
    prototype"; "no certified audit"; prompt-injection risk) which are the most useful thing in it for an agent
    deciding whether to use the service.
- file: trustboost-dev-AGENTS.md
  name: trustboost-agents-md
  method: searched
  source: https://raw.githubusercontent.com/teodorofodocrispin-cmyk/TrustBoost-PII-Sanitizer/main/AGENTS.md
  http_status: 200
  fetched: '2026-09-19'
  description: Provider-published AGENTS.md — when to call TrustBoost, the trial and paid request shapes, the response format, the preflight and verify steps, and the payment details.
  operations: 3
  operation_ids: [sanitize_pii, sanitize_preview, verify_proof]
- file: trustboost-dev-verify-and-score.md
  name: trustboost-dev-verify-and-score
  method: generated
  description: Check a wallet's TrustBoost Score, verify a Proof of Sanitization on Solana, and read a privacy budget — the free read-side operations, with the observed responses for unknown ids.
  operations: 4
  operation_ids: [get_trustboost_score, verify_proof, get_budget_status, health_check]