CanaryCAUTION

The most comprehensive Claude Code guide: agentic workflows, hooks, skills, MCP servers, quizzes, and production-ready templates. 430K+ lines.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: canary
description: "Post-deploy monitoring: watch production after a deploy and alert on regressions"
argument-hint: "[--baseline]"
effort: medium
disable-model-invocation: true
---

# Canary: Post-Deploy monitoring

Watch a live application after deployment. Alert on errors and regressions. Compare against a pre-deploy baseline.

**Two modes:**
- `--baseline`: capture the current state BEFORE deploying
- *(default)*: monitor AFTER deploying and compare against baseline

## Instructions

### Phase 1: Setup

Parse the user's arguments and detect the deployment context.

```bash
# Detect current branch and recent deploy commit
git branch --show-current
git log --oneline -5

# Auto-detect platform from config files
[ -f fly.toml ]         && echo "PLATFORM: fly"
[ -f render.yaml ]      && echo "PLATFORM: render"
[ -f vercel.json ]      && echo "PLATFORM: vercel"
[ -f netlify.toml ]     && echo "PLATFORM: netlify"
[ -f Procfile ]         && echo "PLATFORM: heroku"
[ -f railway.toml ]     && echo "PLATFORM: railway"

# Check for health endpoint
curl -sf "${URL}/health" -w "\n%{http_code}" 2>/dev/null | tail -1
curl -sf "${URL}/api/health" -w "\n%{http_code}" 2>/dev/null | tail -1
```

Create the working directory:

```bash
mkdir -p .canary/baselines .canary/reports .canary/screenshots
```

---

### Phase 2: Baseline Capture (`--baseline` mode)

Run this BEFORE deploying to capture the current healthy state.

For each page to monitor, record:

1. **HTTP status**: is the page returning 200?
2. **Response time**: how long does it take to load?
3. **Content snapshot**: key text content to detect blank pages later

```bash
# For each page URL
for PAGE_PATH in "/" "/dashboard" "/settings" "/api/health"; do
  SLUG=$(echo "$PAGE_PATH" | tr '/' '_' | tr -d '?&=')
  RESULT=$(curl -sf -o /dev/null -w "%{http_code}|%{time_total}" "${BASE_URL}${PAGE_PATH}" 2>/dev/null)
  STATUS=$(echo "$RESULT" | cut -d'|' -f1)
  TIME_MS=$(echo "$RESULT" | awk -F'|' '{printf "%.0f", $2 * 1000}')
  echo "  ${PAGE_PATH}: HTTP ${STATUS}, ${TIME_MS}ms"
done
```

Save baseline to `.canary/baselines/baseline.json`:

```json
{
  "url": "<base-url>",
  "timestamp": "<ISO-8601>",
  "branch": "<branch-name>",
  "commit": "<git-SHA>",
  "pages": {
    "/": { "status": 200, "time_ms": 450 },
    "/dashboard": { "status": 200, "time_ms": 680 },
    "/api/health": { "status": 200, "time_ms": 45 }
  }
}
```

Then **STOP** and tell the user: "Baseline captured. Deploy your changes, then run `/canary <url>` to monitor."

---

### Phase 3: Page discovery

If no pages were specified, auto-discover pages to monitor.

**From the application:**

```bash
# Check sitemap if available
curl -sf "${URL}/sitemap.xml" 2>/dev/null | grep -oP '(?<=<loc>)[^<]+' | head -10

# Check robots.txt for known paths
curl -sf "${URL}/robots.txt" 2>/dev/null | grep -i "allow\|disallow" | head -10

# Common paths to always check
echo "Always check: / /login /dashboard /settings /api/health"
```

Default pages to monitor if nothing found: `/`, and the homepage only.

---

### Phase 4: Monitoring loop

Monitor for the specified duration (default: 10 minutes). Run a check every 60 seconds.

**Each check cycle:**

```bash
TIMESTAMP=$(date -u +%Y-%m-%dT%H:%M:%SZ)
CHECK_NUM=$((CHECK_NUM + 1))

for PAGE_PATH in "${PAGES[@]}"; do
  # Check HTTP status and response time
  RESULT=$(curl -sf -o /dev/null -w "%{http_code}|%{time_total}" \
    --max-time 10 "${BASE_URL}${PAGE_PATH}" 2>/dev/null || echo "0|0")
  STATUS=$(echo "$RESULT" | cut -d'|' -f1)
  TIME_MS=$(echo "$RESULT" | awk -F'|' '{printf "%.0f", $2 * 1000}')

  # Compare against baseline
  BASELINE_STATUS=$(jq -r ".pages[\"${PAGE_PATH}\"].status // 200" .canary/baselines/baseline.json 2>/dev/null)
  BASELINE_TIME=$(jq -r ".pages[\"${PAGE_PATH}\"].time_ms // 1000" .canary/baselines/baseline.json 2>/dev/null)

  echo "  [Check #${CHECK_NUM}] ${PAGE_PATH}: HTTP ${STATUS} (${TIME_MS}ms)"
done
```

**Alert levels:**

| Level | Condition | Trigger |
|-------|-----------|---------|
| **CRITICAL** | Page load failure | HTTP status is not 2xx, curl timeout, DNS failure |
| **HIGH** | New errors | Error rate increased vs baseline (console errors, 5xx responses) |
| **MEDIUM** | Performance regression | Response time exceeds 2x baseline |
| **LOW** | New broken links | Previously-working routes now return 404 |

**Key principles:**
- **Alert on changes, not absolutes.** A page with 3 errors in baseline is fine if still 3. One NEW error is an alert.
- **Transient tolerance.** Only alert on patterns persisting across 2+ consecutive checks. A single network blip is not an alert.

**When a CRITICAL or HIGH alert fires (2 consecutive checks):**

```
CANARY ALERT
=====================================
Time:     [check #N at Xs elapsed]
Page:     [URL]
Level:    [CRITICAL / HIGH / MEDIUM / LOW]
Finding:  [what changed, be specific]
Baseline: [baseline value]
Current:  [current value]
=====================================
Options:
  A) Investigate now: stop monitoring, focus on this issue
  B) Continue monitoring: wait for next check to confirm
  C) Rollback: revert the deploy
  D) Dismiss: known issue, continue monitoring
```

---

### Phase 5: Health report

After monitoring completes (or user stops), produce a summary.

```
CANARY REPORT: [url]
=========================================
Duration:    [X minutes]
Checks:      [N total per page]
Pages:       [N pages monitored]
Commit:      [deployed SHA]
Status:      [HEALTHY / DEGRADED / BROKEN]

Per-Page Results:
-----------------------------------------
  Page           Status      Avg Time   Alerts
  /              HEALTHY     450ms      0
  /dashboard     DEGRADED    1100ms     1 medium (was 450ms)
  /settings      HEALTHY     380ms      0
  /api/health    HEALTHY     45ms       0

Alerts Fired: [N] (X critical, Y high, Z medium, W low)

VERDICT: [DEPLOY HEALTHY / DEPLOY HAS ISSUES (see alerts above)]
=========================================
```

Save report to `.canary

CAUTIONgrade B · trust 89/100 Install with care. The audit found things worth knowing before you trust its output.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (2)

whitepapers/recap-cards/en/_extensions

whitepapers/recap-cards/fr/_extensions

Gates applied: no_behavioural_pass.