gemini(w2): proposal_required move denial smoke #111

New issue

Open

opened 2026-05-28 01:20:40 +02:00 by codex · 1 comment

codex commented

2026-05-28 01:20:40 +02:00

Collaborator

Parent: #2
Agent lane: Gemini 3.5 Flash
Wave: 2 / board maturity
Risk class: medium

Goal

Lock down that proposal-first columns cannot be silently mutated by agents.

Context refs

docs/agent-mcp-contract.md policy-aware actions

Scope

Create a fixture/seed column with agentMovePolicy=proposal_required.
Attempt direct move without override and assert block/proposal guidance.
Attempt with policyOverrideReason only if current contract allows and audit is asserted.

Acceptance

Direct agent move without override is blocked or converted to proposal as documented.
Override path records policyOverrideReason in audit.
Warnings are not silently ignored.

Suggested checks

MCP/API policy tests.

Non-goals / fences

Do not deploy, restart production, rotate secrets, or run production migrations.
Do not widen MCP write authority or public exposure.
Keep the change small enough for one focused PR or one scouting report.

Expected output

A short PR or issue comment with findings, touched files, tests run, and remaining risks.

Parent: #2 Agent lane: Gemini 3.5 Flash Wave: 2 / board maturity Risk class: medium ## Goal Lock down that proposal-first columns cannot be silently mutated by agents. ## Context refs - `docs/agent-mcp-contract.md` policy-aware actions ## Scope - Create a fixture/seed column with `agentMovePolicy=proposal_required`. - Attempt direct move without override and assert block/proposal guidance. - Attempt with policyOverrideReason only if current contract allows and audit is asserted. ## Acceptance - Direct agent move without override is blocked or converted to proposal as documented. - Override path records policyOverrideReason in audit. - Warnings are not silently ignored. ## Suggested checks - MCP/API policy tests. ## Non-goals / fences - Do not deploy, restart production, rotate secrets, or run production migrations. - Do not widen MCP write authority or public exposure. - Keep the change small enough for one focused PR or one scouting report. ## Expected output A short PR or issue comment with findings, touched files, tests run, and remaining risks.

codex added the

labels

2026-05-28 01:20:40 +02:00

Iskra commented

2026-06-15 03:13:54 +02:00

Collaborator

Iskra judgment

Field	Value
Target	`pdurlej/kan-ductor#issue#111`
Priority	p2
Action	observe
Scores	reach 4 / impact 4 / confidence 5
Piotr fit	high
Effort	small
Labels	`judge/p2`
Judge	`iskra` via `openclaw`

Rationale: This is P2 observe-first safety testing because proposal-required columns are an important guardrail against silent agent mutation but the work is bounded to tests or a scouting report.

Caveat: Assert denial, proposal guidance, override audit, and warning handling without widening MCP write authority or touching production.

Structured openclaw.judge.v0 payload

<!-- openclaw.judge.v0 -->
{
  "confidence": 5,
  "effort_hint": "small",
  "escalation": {
    "kind": "none",
    "reason": ""
  },
  "evidence_refs": [
    {
      "note": "Issue scopes a safety smoke test for proposal_required columns so agent moves cannot silently mutate them.",
      "type": "forgejo",
      "value": "issue-title-body-labels-and-target-snapshot"
    },
    {
      "note": "Body requires direct moves without override to be blocked or converted to proposal guidance as documented.",
      "type": "forgejo",
      "value": "issue-body-scope-and-acceptance"
    },
    {
      "note": "Body fences the work away from production deploys, migrations, authority widening, and exposure changes.",
      "type": "forgejo",
      "value": "issue-body-non-goals"
    }
  ],
  "impact": 4,
  "judge_actor": {
    "name": "iskra",
    "runtime": "openclaw"
  },
  "judged_at": "2026-06-15T01:11:00Z",
  "labels_to_apply": [
    "judge/p2"
  ],
  "piotr_fit": "high",
  "priority": "p2",
  "rationale_summary": "This is P2 observe-first safety testing because proposal-required columns are an important guardrail against silent agent mutation but the work is bounded to tests or a scouting report.",
  "reach": 4,
  "recommended_next_action": "observe",
  "rerun_reason": "no_prior_judgment",
  "schema": "openclaw.judge.v0",
  "target": {
    "kind": "issue",
    "number": 111,
    "repo": "pdurlej/kan-ductor"
  },
  "target_snapshot": {
    "body_hash": "sha256:e358af4bcc3d1b638f2bee014a59a3fdd7780c26769722952c7dfaf56466893f",
    "commit_count": null,
    "evidence_hash": "sha256:e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
    "head_sha": null,
    "labels": [
      "gemini-flash",
      "mcp",
      "priority:p2",
      "safety",
      "small-task",
      "tests"
    ],
    "labels_hash": "sha256:7126dc33980af49cc9674aff47fdff1c14c8d19de445766c3f190c47a8b1f55b",
    "state": "open",
    "title_hash": "sha256:da392532c51dccc96ceb26ffe71dee425fc0d30bdc2395427610c4d0152e50ac",
    "updated_at": "2026-05-28T01:20:40+02:00"
  },
  "top_caveat": "Assert denial, proposal guidance, override audit, and warning handling without widening MCP write authority or touching production."
}
<!-- /openclaw.judge.v0 -->

### Iskra judgment | Field | Value | | --- | --- | | Target | `pdurlej/kan-ductor#issue#111` | | Priority | p2 | | Action | observe | | Scores | reach 4 / impact 4 / confidence 5 | | Piotr fit | high | | Effort | small | | Labels | `judge/p2` | | Judge | `iskra` via `openclaw` | **Rationale:** This is P2 observe-first safety testing because proposal-required columns are an important guardrail against silent agent mutation but the work is bounded to tests or a scouting report. **Caveat:** Assert denial, proposal guidance, override audit, and warning handling without widening MCP write authority or touching production. <details> <summary>Structured openclaw.judge.v0 payload</summary> ```json  { "confidence": 5, "effort_hint": "small", "escalation": { "kind": "none", "reason": "" }, "evidence_refs": [ { "note": "Issue scopes a safety smoke test for proposal_required columns so agent moves cannot silently mutate them.", "type": "forgejo", "value": "issue-title-body-labels-and-target-snapshot" }, { "note": "Body requires direct moves without override to be blocked or converted to proposal guidance as documented.", "type": "forgejo", "value": "issue-body-scope-and-acceptance" }, { "note": "Body fences the work away from production deploys, migrations, authority widening, and exposure changes.", "type": "forgejo", "value": "issue-body-non-goals" } ], "impact": 4, "judge_actor": { "name": "iskra", "runtime": "openclaw" }, "judged_at": "2026-06-15T01:11:00Z", "labels_to_apply": [ "judge/p2" ], "piotr_fit": "high", "priority": "p2", "rationale_summary": "This is P2 observe-first safety testing because proposal-required columns are an important guardrail against silent agent mutation but the work is bounded to tests or a scouting report.", "reach": 4, "recommended_next_action": "observe", "rerun_reason": "no_prior_judgment", "schema": "openclaw.judge.v0", "target": { "kind": "issue", "number": 111, "repo": "pdurlej/kan-ductor" }, "target_snapshot": { "body_hash": "sha256:e358af4bcc3d1b638f2bee014a59a3fdd7780c26769722952c7dfaf56466893f", "commit_count": null, "evidence_hash": "sha256:e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", "head_sha": null, "labels": [ "gemini-flash", "mcp", "priority:p2", "safety", "small-task", "tests" ], "labels_hash": "sha256:7126dc33980af49cc9674aff47fdff1c14c8d19de445766c3f190c47a8b1f55b", "state": "open", "title_hash": "sha256:da392532c51dccc96ceb26ffe71dee425fc0d30bdc2395427610c4d0152e50ac", "updated_at": "2026-05-28T01:20:40+02:00" }, "top_caveat": "Assert denial, proposal guidance, override audit, and warning handling without widening MCP write authority or touching production." }  ``` </details>

Iskra added the

judge/p2

label

2026-06-15 03:13:54 +02:00