crawlio-browser 1.5.9 → 1.6.1

package/skills/monitor/SKILL.md

@@ -0,0 +1,66 @@
+---
+name: monitor
+description: "Monitor a URL for changes — baseline capture, re-capture, diff report"
+allowed-tools: Agent
+argument-hint: <url>
+context: fork
+agent: crawlio-investigator
+---
+
+# Monitor Investigation
+
+You are running a **monitor** investigation. Your goal is to capture a baseline snapshot of the target URL, re-capture it, and produce a `DiffReport` showing what changed between the two captures.
+
+## Loop Definition
+
+Read `loops/monitor.json` to understand the phase sequence. The monitor loop has 3 phases:
+
+1. **baseline** — Spawn `crawlio-crawler` to capture the target URL. Record the `EVIDENCE_ID`.
+2. **recapture** — Spawn `crawlio-crawler` again with the same URL. Record the second `EVIDENCE_ID`.
+3. **diff** — Spawn `crawlio-differ` with both evidence IDs. It reads both `EvidenceEnvelope<PageEvidence>` files, compares them field by field across 10 dimensions, and writes an `EvidenceEnvelope<DiffReport>`.
+
+## Execution
+
+1. Read `loops/monitor.json` to confirm phase order.
+2. Spawn `crawlio-crawler` for the baseline capture:
+   ```
+   Crawl <url> and write PageEvidence to .crawlio/evidence/.
+   ```
+   Record `EVIDENCE_ID=<baselineId>`.
+
+3. Spawn `crawlio-crawler` for the recapture:
+   ```
+   Crawl <url> and write PageEvidence to .crawlio/evidence/.
+   ```
+   Record `EVIDENCE_ID=<currentId>`.
+
+4. Spawn `crawlio-differ` with both evidence IDs:
+   ```
+   Compare baseline evidence <baselineId> against current evidence <currentId> for <url>.
+   Read both from .crawlio/evidence/. Write EvidenceEnvelope<DiffReport> to .crawlio/evidence/.
+   ```
+   Record `EVIDENCE_ID=<diffId>`.
+
+5. Read the DiffReport evidence and summarize changes for the user.
+
+## Output Format
+
+```
+## Monitor: <url>
+
+### Changes Detected
+- [List of DiffChange entries grouped by dimension]
+
+### Summary
+- Total changes: N
+- Breaking changes: N
+- Dimensions affected: [list]
+
+### Evidence Chain
+- Baseline: <baselineId> (quality: ...)
+- Current: <currentId> (quality: ...)
+- Diff: <diffId> (quality: ...)
+
+### Confidence
+- Overall: high/medium/low
+```

package/skills/test/SKILL.md

@@ -0,0 +1,103 @@
+---
+name: test
+description: "Automated testing — accessibility, performance, security, SEO, and best-practices audits with testable flow discovery"
+allowed-tools: Agent
+argument-hint: <url>
+context: fork
+agent: crawlio-investigator
+---
+
+# Test Investigation
+
+You are running a **test** investigation. Your goal is to audit a target URL across accessibility, performance, security, SEO, and best-practices categories, discover testable user flows, and produce a `TestSuite` evidence report.
+
+## Loop Definition
+
+Read `loops/test.json` to understand the phase sequence. The test loop has 5 phases:
+
+1. **crawl** — Spawn `crawlio-crawler` to capture the target URL. Record the `EVIDENCE_ID`.
+2. **analyze** — Spawn `crawlio-analyzer` with the crawl evidence ID. Identifies framework and rendering mode to inform test strategy.
+3. **audit** — Spawn `crawlio-auditor` with the crawl evidence ID. Runs accessibility, performance, security, SEO, and best-practices checks.
+4. **discover-flows** (optional) — Spawn `crawlio-auditor` with the crawl evidence ID and flow discovery focus. Discovers testable user flows from navigation/forms.
+5. **synthesize** — Spawn `crawlio-synthesizer` with all phase evidence to produce the final `TestSuite`.
+
+## Execution
+
+1. Read `loops/test.json` to confirm phase order.
+2. Parse the user's argument: `<url>`.
+3. Spawn `crawlio-crawler` to capture the page:
+   ```
+   Crawl <url> and write PageEvidence to .crawlio/evidence/.
+   ```
+   Record `EVIDENCE_ID=<crawlId>`.
+
+4. Spawn `crawlio-analyzer` with the crawl evidence:
+   ```
+   Read PageEvidence from .crawlio/evidence/<crawlId>.json.
+   Analyze framework, rendering mode, and component patterns.
+   Write FrameworkEvidence to .crawlio/evidence/.
+   Target URL: <url>
+   ```
+   Record `EVIDENCE_ID=<analyzeId>`.
+
+5. Spawn `crawlio-auditor` to run audits:
+   ```
+   Read PageEvidence from .crawlio/evidence/<crawlId>.json.
+   Run accessibility, performance, security, SEO, and best-practices audits.
+   Discover testable user flows from navigation and forms.
+   Write TestSuite evidence to .crawlio/evidence/.
+   Target URL: <url>
+   ```
+   Record `EVIDENCE_ID=<auditId>`.
+
+6. Spawn `crawlio-synthesizer` to produce the final TestSuite:
+   ```
+   Read all evidence: <crawlId>, <analyzeId>, <auditId>.
+   Produce a TestSuite with audit results, discovered flows, and overall score.
+   Write to .crawlio/evidence/.
+   Target URL: <url>
+   ```
+   Record `EVIDENCE_ID=<suiteId>`.
+
+7. Read the TestSuite evidence and summarize results for the user.
+
+## Output Format
+
+```
+## Test: <url>
+
+### Summary
+- Score: [0-100]/100
+- Tests: [passed] passed, [failed] failed, [warnings] warnings
+- Flows: [count] testable flows discovered
+
+### Accessibility
+- [audit results with pass/fail/warning status]
+
+### Performance
+- [audit results with scores]
+
+### Security
+- [audit results with pass/fail status]
+
+### SEO
+- [audit results]
+
+### Best Practices
+- [audit results]
+
+### Discovered Flows
+- [list of testable flows with steps]
+
+### Evidence Chain
+- Crawler: <crawlId> (quality: ...)
+- Analyzer: <analyzeId> (quality: ...)
+- Auditor: <auditId> (quality: ...)
+- Suite: <suiteId> (quality: ...)
+
+### Coverage Gaps
+- [Any gaps from the investigation]
+
+### Confidence
+- Overall: high/medium/low
+```