diff --git a/.github/workflows/openclaw-release-telegram-qa.yml b/.github/workflows/openclaw-release-telegram-qa.yml
index e8b0fc01bc57..ddb660287923 100644
--- a/.github/workflows/openclaw-release-telegram-qa.yml
+++ b/.github/workflows/openclaw-release-telegram-qa.yml
@@ -2170,6 +2170,7 @@ jobs:
         env:
           BOUNDARY_EVIDENCE_DIR: ${{ steps.create_sut.outputs.boundary_evidence_dir }}
           OUTPUT_DIR: ${{ steps.run_lane.outputs.output_dir }}
+          RUN_LANE_OUTCOME: ${{ steps.run_lane.outcome }}
         shell: bash
         run: |
           set -euo pipefail
@@ -2213,6 +2214,7 @@ jobs:
               (.candidateArtifact.version | length > 0)
             ' "$context_path" >/dev/null
           jq -e \
+            --arg runLaneOutcome "$RUN_LANE_OUTCOME" \
             '
               .version == 1 and
               .kind == "qa-gateway-process-boundary" and
@@ -2243,7 +2245,10 @@ jobs:
                   )
                 )
               ) and
-              ([.launches[] | select(.terminalState == "ready-exited")] | length >= 2)
+              (
+                [.launches[] | select(.terminalState == "ready-exited")] | length >=
+                  (if $runLaneOutcome == "success" then 2 else 1 end)
+              )
             ' "$runtime_path" >/dev/null
 
           while IFS=$'\t' read -r relative_path expected_sha256; do
diff --git a/AGENTS.md b/AGENTS.md
index 4a8f71333c18..ae93e2d88b4a 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -170,6 +170,9 @@ Skills own workflows; root owns hard policy and routing.
 - `scripts/pr` subcommands require a PR number; no subcommand `--help` placeholder.
 - `scripts/pr` review: checkout main baseline, then PR, before artifact validation.
 - `rg`: options before `--`; use `--` before patterns starting with `-`.
+- `gh --jq` is not standalone `jq`; pipe JSON to `jq` for variables or `--arg`.
+- Actions checkout refs: use full 40-char SHAs; short SHAs resolve as branches/tags.
+- zsh Git object paths: use `${sha}:path`; `$sha:path` invokes parameter modifiers.
 - Bare issue/PR URL/number: inspect live and take the efficient maintainer path; switch branches/refs when useful.
 - No unsolicited PR labels/retitles/rebases/fixups/landing. Comments/reviews ok only for reviewable findings, pre-merge proof, or close/duplicate reason after explicit close/sweep/landing request.
 - Maintainer decision closes the cluster: if deciding reported behavior/proposed fix is not planned, comment+close all directly associated open issues/PRs unless explicitly told to keep one open. Associated means linked PRs/issues, duplicates, companion workaround PRs, and the canonical issue for the rejected behavior.
diff --git a/test/scripts/openclaw-release-telegram-qa-workflow.test.ts b/test/scripts/openclaw-release-telegram-qa-workflow.test.ts
index 6c80ef484cd3..e32a6488ea3d 100644
--- a/test/scripts/openclaw-release-telegram-qa-workflow.test.ts
+++ b/test/scripts/openclaw-release-telegram-qa-workflow.test.ts
@@ -550,6 +550,13 @@ describe("release Telegram QA workflow", () => {
     expect(
       runStep?.run?.indexOf("run_qa_attempt preflight --scenario channel-canary"),
     ).toBeLessThan(runStep?.run?.indexOf("for attempt in 1 2") ?? -1);
+
+    const finalizeStep = job?.steps?.find(
+      (step) => step.name === "Finalize trusted Telegram process-boundary evidence",
+    );
+    expect(finalizeStep?.env?.RUN_LANE_OUTCOME).toBe("${{ steps.run_lane.outcome }}");
+    expect(finalizeStep?.run).toContain('--arg runLaneOutcome "$RUN_LANE_OUTCOME"');
+    expect(finalizeStep?.run).toContain('if $runLaneOutcome == "success" then 2 else 1 end');
   });
 
   it("serializes stderr behind the workflow-command pause", () => {