Fix issue triage workflow by clarifying device log requirements and improving JSON response handling

This commit is contained in:
Ben Meadors
2026-01-30 06:32:08 -06:00
parent 22617076f8
commit 68733a6c51

View File

@@ -100,7 +100,12 @@ jobs:
prompt: |
Analyze this GitHub issue for completeness and determine if it needs labels.
If this looks like a bug on the device/firmware (crash, reboot, lockup, radio issues, GPS issues, display issues, power/sleep issues), request device logs and explain how to get them:
IMPORTANT: Distinguish between:
- Device/firmware bugs (crashes, reboots, lockups, radio/GPS/display/power issues) - these need device logs
- Build/release/packaging issues (missing files, CI failures, download problems) - these do NOT need device logs
- Documentation or website issues - these do NOT need device logs
If this is a device/firmware bug, request device logs and explain how to get them:
Web Flasher logs:
- Go to https://flasher.meshtastic.org
@@ -113,20 +118,18 @@ jobs:
Also request key context if missing: device model/variant, firmware version, region, steps to reproduce, expected vs actual.
Respond ONLY with JSON:
{
"complete": true|false,
"comment": "Your helpful comment requesting missing info, or empty string if complete",
"label": "needs-logs" | "needs-info" | "none"
}
Respond ONLY with valid JSON (no markdown, no code fences):
{"complete": true, "comment": "", "label": "none"}
OR
{"complete": false, "comment": "Your helpful comment", "label": "needs-logs"}
Use "needs-logs" if this is a device bug AND no logs are attached.
Use "needs-logs" ONLY if this is a device/firmware bug AND no logs are attached.
Use "needs-info" if basic info like firmware version or steps to reproduce are missing.
Use "none" if the issue is complete or is a feature request.
Use "none" if the issue is complete, is a feature request, or is a build/CI/packaging issue.
Title: ${{ github.event.issue.title }}
Body: ${{ github.event.issue.body }}
system-prompt: You are a helpful assistant that triages GitHub issues. Be conservative with labels.
system-prompt: You are a helpful assistant that triages GitHub issues. Be conservative with labels. Only request device logs for actual device/firmware bugs, not for build/release/CI issues.
model: openai/gpt-4o-mini
- name: Process analysis result
@@ -137,9 +140,12 @@ jobs:
AI_RESPONSE: ${{ steps.analysis.outputs.response }}
with:
script: |
const raw = (process.env.AI_RESPONSE || '').trim();
let raw = (process.env.AI_RESPONSE || '').trim();
let complete = false;
// Strip markdown code fences if present
raw = raw.replace(/^```(?:json)?\s*/i, '').replace(/\s*```$/i, '').trim();
let complete = true;
let comment = '';
let label = 'none';
@@ -149,9 +155,10 @@ jobs:
comment = (parsed.comment ?? '').toString().trim();
label = (parsed.label ?? 'none').toString().trim().toLowerCase();
} catch {
// If JSON parse fails, treat as incomplete with raw response as comment
complete = false;
comment = raw;
// If JSON parse fails, log warning and don't comment (avoid posting raw JSON)
console.log('Failed to parse AI response as JSON:', raw);
complete = true;
comment = '';
label = 'none';
}
@@ -159,7 +166,9 @@ jobs:
const allowedLabels = new Set(['needs-logs', 'needs-info', 'none']);
if (!allowedLabels.has(label)) label = 'none';
core.setOutput('should_comment', (!complete && comment.length > 0) ? 'true' : 'false');
// Only comment if we have a valid parsed comment (not raw JSON)
const shouldComment = !complete && comment.length > 0 && !comment.startsWith('{');
core.setOutput('should_comment', shouldComment ? 'true' : 'false');
core.setOutput('comment_body', comment);
core.setOutput('label', label);