Entitlement security - 9 tools - one MCP at mcp.atomadic.tech

from atomadic import Atomadic, security
ato = Atomadic(api_key='ato_...')

In plain English - Checks and scores security bubble.

Technical - Bubble check: sandbox-scan a content snippet for adversarial / prompt-injection / exfiltration / jailbreak patterns; returns PROCEED / REVIEW / BLOCK with the threats detected. (plan: pro)

security.assess_security_bubble_pure(ato, content=...)

In plain English - Sorts into categories error fold.

Technical - Classify an error-fold entry with severity, category, and a suggested repair path before Healer ingests it. (plan: dev)

security.classify_error_fold_pure(ato, error_message=...)

In plain English - Calculates hardening posture.

Technical - Compute the recommended hardening posture (cumulative directives) for a target product at a level (info/low/medium/high/critical); critical requires operator co-sign. (plan: pro)

security.compute_hardening_posture_pure(ato, hardening_level=..., target_product_id=...)

In plain English - Calculates redacted args.

Technical - Redact a tool-args object: replaces secret-pattern values and secret-named fields with safe placeholders (nested-aware). (plan: pro)

security.compute_redacted_args_pure(ato, args=...)

In plain English - Calculates redacted text.

Technical - Redact secrets from free-form text; replaces each match with a verify-without-reveal placeholder [REDACTED:kind:sha8]. (plan: pro)

security.compute_redacted_text_pure(ato, text=...)

In plain English - Calculates threat model.

Technical - STRIDE threat model for a component (sample threat + mitigation per category), prioritized by exposure. (plan: pro)

security.compute_threat_model_pure(ato, component=...)

In plain English - Returns the canonical security constants.

Technical - Canonical Security constants: bubble verdicts, hardening levels + cumulative directives, error-fold categories, redaction kinds, PQC standard (FIPS-203). (plan: free)

security.define_security_constants_pure(ato)

In plain English - Scans for dependency risk.

Technical - Scan a requirements list for supply-chain risk (unpinned, wildcard, VCS/URL install, insecure scheme) → findings + verdict. (plan: dev)

security.scan_dependency_risk_pure(ato, requirements_text=...)

In plain English - Checks whether it is valid secret hygiene.

Technical - Scan source for hardcoded secrets (AWS keys, private keys, API tokens, passwords, connection strings) → redacted findings + hygiene verdict. (plan: dev)

security.validate_secret_hygiene_pure(ato, source_text=...)