BF trustboost-pii-sanitizer
Context-aware PII sanitization for autonomous AI agent pipelines. Sanitizes text before LLMs with 5 context modes (legal/financial/medical/code/general), Privacy Budget per agent, and TrustBoost Score for M2M trust verification. Supports EN, ES (LATAM), PT (BR/PT), DE, JA, FR, IT, KO with country-specific patterns (RFC, CUIT, CPF, CNPJ, Personalausweis, マイナンバー, NIR, Codice Fiscale, 주민등록번호). Returns sanitized text, safety_score (0.0-1.0), risk_category (CRITICAL/PRIVATE/SENSITIVE/CLEAN), and context_applied. No SDK required — single POST request. 50 free requests per wallet with tx_hash="TRIAL".
As a process F 33/100 · Will not run — References files that are not bundled: PRIVACY.md
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The purpose itself is risky: wallets, browser password stores, offensive security. Even an honest implementation gives the agent access to things that cost money.
Explain in the description why the access is needed and how it is limited; add tests that show refusals on dangerous requests.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The text contains phrases like "ignore previous instructions" or "you are now…". That is an attempt to hijack the agent: it may break your rules, the system limits or company policy.
An honest skill does not need them: state the role and the rules directly without overriding other instructions. Otherwise catalog scanners and corporate filters will block the listing.
How to improve
- The text references files that are not there: add them or drop the references.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 6
✓ No critical or high findings
Medium and low: 6
-
medium Risky intent
intent-wallet-secretsSKILL.md:78Handles crypto-wallet secrets (seed / mnemonic / private key) — a classic stealer target2. Agent signs with its OWN wallet private key
-
medium Instruction override
en-ignore-previousSKILL.md:401Instruction-override phrase ("ignore previous instructions") (detector / deny-list definition)- **Prompt injection risk:** Malicious text containing instructions like "Ignore previous instructions" could potentially bypass PII redaction. temperature=0 and strict JSON-only output reduce this ri
detector -
low Secrets in code
secret-high-entropy-tokenSKILL.md:103High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)address="giu4…WE4"
quoted -
low Secrets in code
secret-high-entropy-tokenSKILL.md:234High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)`giu4…WE4`
quoted -
low Secrets in code
secret-high-entropy-tokenSKILL.md:280High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)"payment_address": "giu4…WE4"
quoted -
low Secrets in code
secret-high-entropy-tokenSKILL.md:306High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)"payment_address": "giu4…WE4"
quoted
Files scanned: 2. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
missing-refreference to a missing file: PRIVACY.md
Process rating: all ten parameters 33/100
- 0Tools and files. 1 referenced file(s) missing: PRIVACY.md
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Failures and branches. Linear process with no failure handling
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 5 mutating operations with no state check
- 100Steps. 43 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 3018 tokens
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 19 top-level sections: this looks like several domains in one skill
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +2Single-language instructions
- +3Description length 601: enough signal without eating the budget
- +4Structure: 24 headings
- +3Step-by-step instructions: 43 items
- +3Output format is stated explicitly
- +4Has examples (13 code blocks)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 82.