AGENT SKILL

qualify-entropy-source

Capture reproducible entropy-source qualification evidence with source physics, boundaries, estimators, health tests, and artifact hashes.

License 许可
MIT
所属仓库
dasobral/skills
最近更新
2026-08-15
Stars
★ 0
Forks
0

01 前言 Front Matter

name
qualify-entropy-source
description
Capture reproducible entropy-source qualification evidence with source physics, boundaries, estimators, health tests, and artifact hashes.

02 资源文件 8

03 SKILL.md 正文

Qualify Entropy Source

  1. Read references/authoritative-sources.md and all three JSON schemas.
  2. Identify source physics, adversarial assumptions, raw sampling point, digitization, conditioning, hardware, firmware, driver, sampling rate, and operating envelope.
  3. Pin restart matrix identity, estimator versions and invocation argument/hash/exit-code provenance, health-test parameters, and artifact hashes.
  4. Derive RCT/APT parameters with python3 scripts/derive_health_tests.py; record formula version, 0 < alpha < 1, alphabet size, claimed entropy, and the enforced APT window (1024 for a binary alphabet, 512 otherwise).
  5. Run python3 scripts/qualify_run.py < input.json > qualification.json; it recomputes the parameters, rejects mismatches, and records fail when observed estimator min-entropy is below the source's claim.
  6. Treat missing tools or evidence as evidence-gap; require expert review of the resulting evidence.

For native hooks, register a baseline conforming to references/entropy-baseline.schema.json under external PLUGIN_DATA/entropy-flight-recorder (or the explicit plugin state directory). Approval requires a detached Ed25519 signature verified against the public key selected by ENTROPY_FLIGHT_RECORDER_TRUST_ROOT inside the explicit ENTROPY_FLIGHT_RECORDER_TRUST_BOUNDARY. Trust-path components must be non-symlink, root- or ENTROPY_FLIGHT_RECORDER_TRUSTED_UID-owned, non-writable by group/other, and non-replaceable by the invoking unprivileged process. The signed payload binds plugin, decision, approver identity, timestamp, baseline digest, and prior ledger entry. The local HMAC chain is only tamper-evident storage after signature verification; it is not approval authority.

Evidence states are exactly pass, fail, unknown, not-applicable, and evidence-gap. Keep them separate from the review-required decision.

Use authorized test captures only. Never collect secrets, seed material, raw production key material, or raw production entropy streams. This workflow records evidence and does not certify FIPS status, quantum origin, entropy adequacy, compliance, or fitness for use.

04 同仓库更多 Skills

dasobral
dasobral/skills

fixture-skill

Fixture skill used to prove trust inventory coverage.

Skill与智能体开发
★ 0
dasobral
dasobral/skills

write-conformant-code

Write code in any language that matches project conventions exactly. Reads CODING_REQUIREMENTS.md, prior review output, or built-in defaults before writing. Part of Codecraft pipeline: analyze → write → review → audit. Use when implementing features, functions, or any code that should fit seamlessly into an existing codebase.

开发工程师内容创作者写作与文案
★ 0
dasobral
dasobral/skills

test-crypto-interoperability

Record reproducible cryptographic interoperability cases, including negotiation, fallback, sizes, latency, fragmentation, and negative tests.

测试与质量
★ 0
dasobral
dasobral/skills

scientific-platform-architect

Design scientific computing platforms combining HPC kernels with modern infrastructure. Enforces compute-layer separation (Rust/PyTorch/Python), emulator-first thinking, and right-sized infrastructure. Use for MCMC pipelines, neural emulators, platform design, or scientific code review.

设计师代码审查视觉设计
★ 0
dasobral
dasobral/skills

rust-systems

Rust systems engineering for daemons, HTTP APIs, and AOS components. Enforces Tokio/Axum async-first, LiteLLM-compatible OpenAI APIs, workspace layouts, tracing, and systemd-compatible daemon design. Part of AOS Stack plugin. Use for .rs files, Cargo.toml, daemons, or Rust orchestrator services.

视觉设计
★ 0
dasobral
dasobral/skills

run-agent-attack-replay

Aggregate isolated agent attack trials with deterministic side-effect assertions, Wilson intervals, and preserved trace evidence.

Skill与智能体开发
★ 0