Expand description
SDK-free skill manifests, loaders, registries, and verification contracts.
Structs§
- Asset
- Inert asset recorded during activation.
- Budgets
- Skill-requested budget limits.
- Deny
Verifier DenyVerifieris the canonical operator-deny verifier: it rejects every manifest unconditionally. Use it as a safe default until a real signature verifier (e.g., Sigstore/cosign) adapter is wired in.- Effective
Envelope - Minimal effective envelope result used by activation code.
- Human
Approval Step - Human approval checkpoint.
- InMemory
Registry - In-memory registry implementation.
- Loader
- Filesystem loader for skill packs.
- Manifest
- Locked skill manifest.
- MaxCost
- Maximum cost budget.
- Model
Hints - Optional model hints.
- Pack
- Loaded skill pack.
- Progressive
Disclosure - Progressive disclosure copy.
- Prompt
Ref - Prompt reference with explicit version.
- References
- References to executable and context-bearing registrations by name/pattern.
- Requires
- Activation preconditions.
- Script
Asset - Inert script asset metadata.
- Signature
- Signature metadata carried by the manifest.
- Skill
Loader Config - Config-loader compatible skill activation source.
- Warn
Only Verifier - Default warn-only verifier.
Enums§
- Safety
- Skill safety order. Informational in manifests; effective safety is computed from tools.
- Skill
Error - Skill errors.
- Verification
Outcome - Verification result.
Constants§
- MANIFEST_
FILE_ NAME - Locked skill manifest filename.
- SKILL_
MD_ FILE_ NAME - Progressive-disclosure body filename.
Traits§
- Provider
- Source of skill packs.
- Registry
- Registry abstraction for skills.
- Verifier
- Verifier for manifest signatures.
Functions§
- effective_
envelope - Compute a conservative effective envelope by intersecting scope sets and maxing safety.
- effective_
safety - Compute effective safety as the maximum over referenced tool envelope safety values.
- register_
provider - Explicitly register all packs from a provider into a registry.