Trust (Delegation)

openkoi trust manages trust and delegation — granting the agent autonomous action in specific domains, revoking anytime, and auditing every decision it made on its own.

Trust is earned, not configured. The agent starts at LOW trust in every domain. As it demonstrates accuracy (visible in openkoi reflect and openkoi mind calibrate), you can grant higher trust levels.

Subcommands

Subcommand	Description
show	Current trust level per domain
grant <domain> <level>	Delegate a domain. Levels: ask, suggest, act, autonomous
revoke <domain>	Revoke delegation for a domain
audit [domain]	Audit autonomous actions taken (optionally filter by domain)

When run without a subcommand, defaults to show.

Trust Levels

Level	Name	Behavior
NONE	Never	Agent will never act in this domain, even if asked
LOW / ask	Always ask	Agent asks before every action
MEDIUM / suggest	Suggest + approve	Agent proposes actions, waits for your approval
HIGH / act	Act, report after	Agent acts autonomously, reports what it did
FULL / autonomous	Fully delegated	Agent acts and only reports in daily reflection

openkoi trust show

$ openkoi trust show

╭─────────────────────────────────────────────────────────────╮
│ TRUST LEVELS                                                 │
│                                                              │
│  Domain               Trust    Mode           Since          │
│  ─────────────────────────────────────────────────────────   │
│  code-review           HIGH    Delegated      47 days ago    │
│  test-generation       HIGH    Delegated      32 days ago    │
│  commit-messages       HIGH    Delegated      28 days ago    │
│  email-drafting        MEDIUM  Suggest+Approve 20 days ago   │
│  slack-replies         MEDIUM  Suggest+Approve 15 days ago   │
│  file-operations       LOW     Always ask      —             │
│  deploy                LOW     Always ask      —             │
│  money/purchases       NONE    Never           —             │
│                                                              │
│ Grant trust:  openkoi trust grant <domain> <level>           │
│ Revoke trust: openkoi trust revoke <domain>                  │
╰─────────────────────────────────────────────────────────────╯

openkoi trust grant

Grant higher trust to a specific domain:

# Grant autonomous code review
openkoi trust grant code-review autonomous

# Grant suggest-level email drafting
openkoi trust grant email-drafting suggest

# Grant basic trust for deployment
openkoi trust grant deploy ask

The agent confirms the grant and explains what it means:

$ openkoi trust grant code-review autonomous

  code-review trust level: LOW → AUTONOMOUS
  The agent will now:
    • Auto-approve low-risk PRs
    • Request changes on issues it detects
    • Flag security concerns for human review
    • Report all actions in daily reflection

  Revoke anytime: openkoi trust revoke code-review

openkoi trust revoke

Immediately revoke delegation for a domain:

openkoi trust revoke code-review

$ openkoi trust revoke code-review

  code-review trust level: AUTONOMOUS → LOW
  The agent will now ask before any code review action.

openkoi trust audit

Review every autonomous action the agent took. This is the accountability layer — every action taken under delegation is logged and auditable:

$ openkoi trust audit

╭─────────────────────────────────────────────────────────────╮
│ AUTONOMOUS ACTION AUDIT — Last 7 days                        │
│                                                              │
│ Domain: code-review (delegated)                              │
│                                                              │
│  Mar 4   Auto-approved PR #147 (test coverage: 94%)         │
│  Mar 3   Requested changes on PR #145 (missing error        │
│          handling). Author fixed in 2hrs.                     │
│  Mar 2   Auto-approved PR #143 (docs update)                │
│  Mar 1   Flagged PR #141 for human review (security)        │
│                                                              │
│ Judgment accuracy: 4/4 (100%)                                │
│ Human overrides: 0                                           │
│ Trust recommendation: MAINTAIN                               │
╰─────────────────────────────────────────────────────────────╯

Filter by domain:

openkoi trust audit code-review
openkoi trust audit email-drafting

Without a domain argument, shows all autonomous actions across all domains.

Trust Recommendations

After each audit, the system provides a recommendation:

Recommendation	Meaning
MAINTAIN	Actions were appropriate. Keep current trust level.
INCREASE	Consistently excellent judgment. Consider granting higher trust.
DECREASE	Multiple human overrides or errors. Consider revoking.
REVOKE	Serious judgment failure. Immediate revocation recommended.

How Trust Influences the System

Trust levels are not just UI — they actively change agent behavior:

System	How Trust Affects It
Guardian	Higher trust = fewer safety blocks in trusted domains
Orchestrator	Autonomous domains skip human confirmation
Daemon	Only acts on domains with HIGH+ trust during background runs
Maturity stages	Stage 3 (Trusted Delegate) requires 3+ HIGH trust domains

Building Trust Over Time

The recommended path from zero to full delegation:

1. Start at LOW — Agent asks before every action
2. After 1-2 weeks: Review openkoi trust audit and openkoi mind calibrate
3. If accuracy > 90%: Grant suggest level — agent proposes, you approve
4. After another 2 weeks: If accuracy stays high, grant act level
5. After a month: If no human overrides, consider autonomous

This mirrors how you'd build trust with a new team member — gradual delegation based on demonstrated competence.

Related Commands

• openkoi reflect today — See decisions and outcomes for trust evaluation
• openkoi mind calibrate — Agency accuracy data supports trust decisions
• openkoi soul show — Trust domains are reflected in the soul
• Security: Trust Levels — How trust fits into the security model