What it is
SM is the release-mediation system for secrets. It proves whether a secret was released, to whom, under what conditions, or whether release was refused.
SM / Truth Stack
Secret Mediation
Controlled secret mediation under proof.
SM is the release-mediation system for secrets. It proves whether a secret was released, to whom, under what conditions, or whether release was refused.
What problem it solves
Secret managers optimize for retrieval and caching. Under coercion, audit, or litigation, that model collapses because disclosure truth is weak and often unverifiable.
What it does
- Accepts sealed secrets under doctrine and binds release to capability, time, identity, and execution context.
- Supports mediation primitives such as lease, use-once, derive, inject-into-process, and transform-without-disclosure.
- Emits release, revocation, and refusal evidence as first-class proof.
What it does not do
- It does not expose a raw plaintext retrieval primitive.
- It does not act as long-term secret storage, a general KV database, or an admin bypass layer.
- It does not privilege convenience, caching, or silent rotation over proof.
Who it is for
- Operators who treat secrets as liabilities to constrain rather than assets to distribute.
- Systems where blast radius and future provability outweigh developer ergonomics.
Where it fits
SM is the Truth Stack mediation layer. It arbitrates disclosure without owning authority or material lifecycle.
Typical deployment context
Used where secret use must remain bounded, re-authorized, and provable years later.
Primary source docs
- Why SMsm/docs/01_why_sm.md
- Refusal doctrinesm/docs/03_refusal_doctrine.md
- CLI/API contractsm/docs/06_cli_api_contract.md