workspace-monitor

name	workspace-monitor
description	Proactive log and build monitoring. Scans application logs, test results, build output, and CI artifacts with smart filtering to detect issues before users report them. Language-aware via CC_LANGUAGE — adapts error patterns per stack.
user-invocable	true
context	fork
disable-model-invocation	true
allowed-tools	Bash, Read, Write, Grep, Glob
argument-hint	--errors-only | --since=1h | --build | --tests | --runtime | --all
featured	true
featured_description	Proactive log and build monitoring that detects issues before you report them.

Workspace Monitor — Proactive Log and Build Intelligence

Scans application logs, test results, build output, and CI artifacts for the current project. Uses smart filtering to detect errors, warnings, stack traces, and anomalies. Language-aware — adapts scanning patterns based on CC_LANGUAGE from cognitive-core.conf.

Operates in two modes:

• Single-project (default): Scans CC_PROJECT_DIR only
• Workspace (opt-in): If CC_WORKSPACE_PROJECTS is defined, scans multiple project directories

Arguments

• $ARGUMENTS — options:
  • --errors-only — only show ERROR/FATAL level entries
  • --since=1h|6h|24h|7d — time window (default: CC_MONITOR_SINCE or 24h)
  • --build — focus on build/compile output
  • --tests — focus on test results
  • --runtime — focus on application runtime logs
  • --tail — show last N lines of active logs
  • --all — scan all categories (default behavior)

Configuration

From cognitive-core.conf:

Variable	Default	Description
CC_LANGUAGE	—	Determines default error patterns and log conventions
CC_MONITOR_LOG_DIRS	logs	Log directories to scan (space-separated, relative to project root)
CC_MONITOR_TEST_DIRS	test-results	Test result directories (space-separated, relative)
CC_MONITOR_BUILD_DIRS	target dist build	Build artifact directories (space-separated, relative)
CC_MONITOR_EXTRA_PATTERNS	—	Additional error regex (pipe-separated, appended to language defaults)
CC_MONITOR_SINCE	24h	Default time window
CC_MONITOR_REPORT_DIR	${CC_SESSION_DOCS_DIR}/monitor-reports	Report output directory
CC_MONITOR_MAX_LOG_SIZE	104857600	Max log file size (bytes) before alerting (100MB)
CC_WORKSPACE_PROJECTS	—	Optional: additional project directories for multi-project scanning (absolute paths, space-separated)

Live Context (auto-injected)

Active Log Files

!for dir in ${CC_MONITOR_LOG_DIRS:-logs}; do [ -d "${dir}" ] && find "${dir}" -name "*.log" -mtime -1 2>/dev/null | head -5; done

Recent Test Results

!for dir in ${CC_MONITOR_TEST_DIRS:-test-results}; do ls -lt "${dir}"/*.xml 2>/dev/null | head -3; ls -lt "${dir}"/*.tap 2>/dev/null | head -3; done

Log File Sizes

!for dir in ${CC_MONITOR_LOG_DIRS:-logs}; do [ -d "${dir}" ] && ls -lh "${dir}"/*.log 2>/dev/null; done

Build Artifacts

!for dir in ${CC_MONITOR_BUILD_DIRS:-target dist build}; do [ -d "${dir}" ] && echo "${dir}: exists ($(find "${dir}" -maxdepth 2 \( -name "*.jar" -o -name "*.war" -o -name "*.whl" -o -name "*.tar.gz" \) 2>/dev/null | wc -l | tr -d ' ') artifacts)"; done

cognitive-core Security Log

![ -f .claude/cognitive-core/security.log ] && tail -10 .claude/cognitive-core/security.log 2>/dev/null || echo "No security log"

Instructions

Step 1: Load Configuration and Patterns

1. Source cognitive-core.conf for CC_LANGUAGE, CC_MONITOR_* variables
2. Load language-specific monitor patterns if available:
   • Check language-packs/${CC_LANGUAGE}/monitor-patterns.conf (in framework source)
   • Or read inline patterns from references/error-patterns.md (in this skill directory)
3. Merge CC_MONITOR_EXTRA_PATTERNS with language defaults (pipe-separated regex)

Step 2: Determine Scope

• Single-project (default): Scan CC_PROJECT_DIR only
• Workspace mode: If CC_WORKSPACE_PROJECTS is defined, scan each listed directory
• Apply $ARGUMENTS to select categories: --build, --tests, --runtime, or --all

Step 3: Discover Log Files

For each project directory in scope:

1. Search configured CC_MONITOR_LOG_DIRS for *.log files
2. Search for additional log locations: **/logs/, **/*.log (limit depth to 3)
3. Check logging configuration files based on CC_LANGUAGE:
   • Perl: logger.conf, environments/*.yml (Log4perl)
   • Java: src/main/resources/logback*.xml, src/main/resources/log4j2.xml
   • Python: logging.conf, logging.yaml, pyproject.toml [tool.logging]
   • Node: winston.config.*, pino config in package.json
4. Note file sizes and last modification times

Step 4: Smart Scan — Runtime Logs

If --runtime or --all:

For each discovered log file:

1. Recency check: Skip files not modified within the --since window
2. Error extraction: Apply language-aware error patterns (see references/error-patterns.md)
3. Context capture: For each match, capture 3 lines before and 3 after
4. Deduplication: Group identical error messages, count occurrences
5. Severity classification:
   • CRITICAL: Unhandled exceptions, OOM, security events, data corruption indicators
   • ERROR: Application errors, failed operations, connection issues
   • WARNING: Deprecations, slow queries, resource pressure, high retry counts
   • INFO: Notable events that aren't errors but may indicate trends

Step 5: Parse Test Results

If --tests or --all:

1. Find test result files in CC_MONITOR_TEST_DIRS:
   • JUnit XML: **/TEST-*.xml, **/junit.xml, **/surefire-reports/*.xml
   • TAP: *.tap, *.t output
   • pytest: **/pytest-report.xml
   • Jest: **/jest-results.json
2. Detect format based on CC_LANGUAGE and file extension
3. Parse test counts: total, passed, failed, skipped, errors
4. Calculate pass rate per test suite
5. If CC_TEST_COMMAND is set, offer to re-run for live results

Step 6: Check Build Health

If --build or --all:

1. Check CC_MONITOR_BUILD_DIRS for build artifacts
2. Detect build system from CC_LANGUAGE:
   • Java: target/ (Maven), build/ (Gradle) — look for BUILD SUCCESS/BUILD FAILURE
   • Python: dist/ (setuptools/poetry), build/ (flit)
   • Node: dist/, build/, .next/ — look for bundle errors
   • Perl: blib/ (ExtUtils), local/ (Carton)
3. Check for stale builds (artifacts older than source files)
4. Verify artifact integrity (exists and non-zero size)
5. Look for build failure indicators in build logs

Step 7: Log Growth Analysis

For all discovered log files:

1. Check current size against CC_MONITOR_MAX_LOG_SIZE
2. If a log exceeds the threshold, flag as alert
3. Check for log rotation configuration (presence of archived logs)
4. Flag missing rotation for large log files

Step 8: Generate Monitoring Report

Save to ${CC_MONITOR_REPORT_DIR:-docs/monitor-reports}/YYYY-MM-DD-monitor.md:

# Monitoring Report

> Generated: YYYY-MM-DD HH:MM
> Project: [CC_PROJECT_NAME] ([CC_LANGUAGE])
> Window: last [since]
> Mode: [single-project | workspace (N projects)]

## Alerts

| Severity | Source | Message | Count | Last Seen |
|----------|--------|---------|-------|-----------|
| CRITICAL | [log file] | [error msg] | [N] | [timestamp] |
| ERROR | [log file] | [error msg] | [N] | [timestamp] |
| WARNING | [log file] | [warning msg] | [N] | [timestamp] |

## Test Results

| Suite | Tests | Passed | Failed | Errors | Skipped | Pass Rate |
|-------|-------|--------|--------|--------|---------|-----------|
| [name] | [N] | [N] | [N] | [N] | [N] | [%] |

## Build Health

| Directory | Last Build | Status | Artifacts | Stale? |
|-----------|-----------|--------|-----------|--------|
| [dir] | [time ago] | [status] | [count] | [yes/no] |

## Log Growth

| File | Size | Threshold | Rotation | Alert |
|------|------|-----------|----------|-------|
| [path] | [size] | [max] | [yes/no] | [OK/WARN] |

## Error Details

### [Source File]

[timestamp] [level] [context] error message with stack trace ... (N occurrences in window)

## Recommended Actions
1. [Priority action based on findings]

Step 9: Workspace Mode — Aggregate Dashboard

When CC_WORKSPACE_PROJECTS is defined and scanning multiple projects:

• Prefix all table rows with the project name
• Add a "Summary by Project" section at the top
• Aggregate alert counts across projects

Examples

/workspace-monitor                                   # Default: full scan, last 24h
/workspace-monitor --errors-only --since=1h          # Errors only, last hour
/workspace-monitor --tests                           # Test results only
/workspace-monitor --build                           # Build health only
/workspace-monitor --runtime --since=7d              # Runtime logs, last week

See Also

• /project-status — Development progress report (git history based)
• /fitness — Codebase fitness scoring (code quality)
• /session-resume — Session context recovery
• references/error-patterns.md — Default error patterns by language
• cognitive-core.conf — All configuration reference