Summary

• Problem: Telegram/native allow-always approvals already persist allowlist entries with source: "allow-always", but the gateway exec.approvals.set schema rejected that field as unexpected.
• Why it matters: local approvals files and the gateway push path disagree on the allowlist contract, so gateway sync can fail even when the on-disk approvals file is valid for the runtime.
• What changed: src/gateway/protocol/schema/exec-approvals.ts now accepts optional source: "allow-always" on allowlist entries, matching the existing persisted approvals type.
• What did NOT change (scope boundary): no approval decision logic, persistence behavior, or Telegram/native reply handling changed in this PR; it only aligns the gateway protocol schema with existing stored data.
• AI-assisted: Yes. I implemented the protocol fix and verified it locally with targeted tests and changed-scope checks.

Change Type (select all)

• Bug fix
• Feature
• Refactor required for the fix
• Docs
• Security hardening
• Chore/infra

Scope (select all touched areas)

• Gateway / orchestration
• Skills / tool execution
• Auth / tokens
• Memory / storage
• Integrations
• API / contracts
• UI / DX
• CI/CD / infra

Linked Issue/PR

• Closes [Bug]: Telegram /approve allow-always writes 'source' field that openclaw approvals set --gateway rejects — allowlist updates silently fail to persist to gateway openclaw/openclaw#69482
• Related [Bug]: enqueueSystemEvent not deduplicated by runId/contextKey — agents cascade duplicate exec approval prompts under new IDs, locking ecosystem openclaw/openclaw#69478
• This PR fixes a bug or regression

Root Cause (if applicable)

• Root cause: the durable approvals store type and normalization path already preserved source: "allow-always", but the gateway protocol schema for exec.approvals.set and exec.approvals.node.set did not allow that field.
• Missing detection / guardrail: there was no validator regression test covering a real allowlist payload that included source: "allow-always".
• Contributing context (if known): the issue report shows this surfaced through Telegram "Allow always", but the mismatch is fundamentally between persisted approvals data and the gateway wire contract.

Regression Test Plan (if applicable)

• Coverage level that should have caught this:
  • Unit test
  • Seam / integration test
  • End-to-end test
  • Existing coverage already sufficient
• Target test or file: src/gateway/protocol/index.test.ts
• Scenario the test should lock in: validateExecApprovalsSetParams and validateExecApprovalsNodeSetParams accept allowlist entries that include source: "allow-always".
• Why this is the smallest reliable guardrail: the bug is a protocol/schema rejection, so validator tests hit the exact contract boundary that failed without needing a live Telegram or gateway sync flow.
• Existing test that already covers this (if any): none.
• If no new test is added, why not: N/A.

User-visible / Behavior Changes

• Gateway approval set payloads may now include persisted allow-always source metadata without being rejected as invalid.

Diagram (if applicable)

Before:
[allowlist entry with source="allow-always"] -> [exec.approvals.set validator] -> [reject unexpected property]

After:
[allowlist entry with source="allow-always"] -> [exec.approvals.set validator] -> [accept payload]

Security Impact (required)

• New permissions/capabilities? (Yes/No) No
• Secrets/tokens handling changed? (Yes/No) No
• New/changed network calls? (Yes/No) No
• Command/tool execution surface changed? (Yes/No) No
• Data access scope changed? (Yes/No) No
• If any Yes, explain risk + mitigation:

Repro + Verification

Environment

• OS: Ubuntu Linux in Cursor Cloud
• Runtime/container: Node 22 + pnpm workspace
• Model/provider: N/A
• Integration/channel (if any): Gateway exec approvals protocol validation
• Relevant config (redacted): approvals payload with allowlist entry { pattern: "/usr/bin/curl", source: "allow-always" }

Steps

1. Construct an approvals file payload containing an allowlist entry with source: "allow-always".
2. Validate it through the gateway protocol path for exec.approvals.set or exec.approvals.node.set.
3. Observe whether the validator accepts or rejects the payload.

Expected

• The gateway protocol should accept allowlist entries that match the already-persisted approvals type, including source: "allow-always".

Actual

• Before this fix, the schema rejected the payload because source was not part of ExecApprovalsAllowlistEntrySchema.

Evidence

• Failing test/log before + passing after
• Trace/log snippets
• Screenshot/recording
• Perf numbers (if relevant)

Human Verification (required)

• Verified scenarios: ran pnpm test src/gateway/protocol/index.test.ts; ran pnpm check:changed.
• Edge cases checked: both gateway set and node set validators accept allow-always source metadata; no unrelated schema fields were broadened.
• What you did not verify: live Telegram approval flow or a real openclaw approvals set --gateway round-trip against a running gateway.

Review Conversations

• I replied to or resolved every bot review conversation I addressed in this PR.
• I left unresolved only the conversations that still need reviewer or maintainer judgment.

Compatibility / Migration

• Backward compatible? (Yes/No) Yes
• Config/env changes? (Yes/No) No
• Migration needed? (Yes/No) No
• If yes, exact upgrade steps:

Risks and Mitigations

• Risk: this only fixes the schema mismatch slice of the issue; if another runtime path strips or mishandles source, that would need follow-up.
  • Mitigation: this PR is intentionally scoped to the concrete validator rejection, and the new tests lock the contract in place.