Gates

Gates are mechanical brakes. Non-zero exit = STOP.

They do not replace judgment; they prevent pretending success when structure or proof is missing.

Gate catalog

Command	Stops when
`retornatus gate contract <C-id>`	No active Contract, empty WHAT, or no DONE criteria
`retornatus gate skill-research <S-id>`	RESEARCH lacks sources / PROCEDURE blank
`retornatus gate evidence <C-id>`	No Evidence artifacts for the Change
`retornatus gate assurance <C-id>`	Assurance is not `SATISFIED`
`retornatus gate policy <A-id>`	Policy is `DENY` or `REQUIRE_HUMAN`
`retornatus verify <C-id>`	Same family as assurance over Contract DONE Claims

retornatus policy check --action <A-id>
retornatus policy check --effect "…"
retornatus run <A-id> --strict-policy

Verdict	Meaning
`SATISFIED`	Each required Claim has bound, fresh, type-appropriate Evidence
`NOT_SATISFIED`	Evidence exists but wrong claim/subject/type or stale
`INCONCLUSIVE`	Required Evidence missing or unbound

> Agent conclusion ≠ Evidence. A green test suite is evidence of tests — not automatic proof of every Claim unless bound correctly.

	Process (hub skill)	Brakes (gates)
Who checks	Agent follows the skill	CLI exit codes
Incomplete work	Agent should stop	Agent cannot claim success via gate
Best for	Learning the method	Teams that need proof between approvals

Some gates can be skipped only as a recorded decision, e.g.:

retornatus skill activate S-0001 --force --reason "Emergency hotfix; research deferred"

Bypass without reason/authority is rejected. See Governance.

Guarantee	Mechanism
Contract before build	`gate contract` + Situation sufficiency
Research before specialized execution	`gate skill-research`
Evidence before done	Claim-bound Evidence + `verify`
Continuity after crash	Canonical files + `wake`
No silent Rule authority	Human Decision required
Policy visibility	`gate policy` / `--strict-policy`