warden logs

View, replay, and manage saved Warden analysis logs. Logs are automatically saved to .warden/logs/ in JSONL format.

Usage

# List all saved runs
warden logs
warden logs list

# Show results from specific log file
warden logs show <file.jsonl>
warden logs show <run-id>

# Clean up expired logs
warden logs gc

Subcommands

list (default)

List all saved analysis runs with summary information.

warden logs
warden logs list

Output columns:

RUN - Short run ID (first 8 characters)
DATE - Relative timestamp (e.g., “2h ago”, “yesterday”)
FILES - Number of files analyzed
FINDINGS - Severity breakdown (high / medium / low)
TIME - Analysis duration
COST - API costs in USD
SHA - Git commit SHA (first 7 characters)
MODEL - Claude model used
SKILLS - Skills executed in the run

Example:

$ warden logs

  RUN       DATE        FILES  FINDINGS      TIME      COST    SHA      MODEL              SKILLS
  a1b2c3d4  2h ago         12  0 / 3 / 1     45s       $0.12   abc1234  claude-sonnet-4.5  security-review, bug-detection
  e5f6g7h8  yesterday      8   2 / 1 / 0     32s       $0.08   def5678  claude-sonnet-4.5  security-review
  i9j0k1l2  2d ago        15   0 / 0 / 5     1m 12s    $0.19   ghi9012  claude-sonnet-4.5  bug-detection, code-quality

3 runs  ·  14 findings  0 / 4 / 6  ·  1m 49s  ·  $0.39  ·  3 skills

show

Display full results from one or more saved log files.

# By file path
warden logs show .warden/logs/run-20260305-143022.jsonl

# By run ID (short or full)
warden logs show a1b2c3d4
warden logs show a1b2c3d4-e5f6-7890-abcd-ef1234567890

# Multiple files
warden logs show run1.jsonl run2.jsonl

# With filtering
warden logs show a1b2c3d4 --report-on high
warden logs show a1b2c3d4 --min-confidence high

The show subcommand is optional when the first argument looks like a file or run ID:

# These are equivalent
warden logs show a1b2c3d4
warden logs a1b2c3d4

gc

Garbage collect (remove) expired log files based on retention policy.

warden logs gc

Retention period is configured in warden.toml (default: 30 days):

[logs]
retentionDays = 30

Options

For list

--json

boolean

default:"false"

Output log list as JSON array

warden logs list --json

For show

--json

boolean

default:"false"

Output results in JSON format

warden logs show a1b2c3d4 --json

--report-on

string

Filter results to show only findings at or above severityValues: off, critical, high, medium, low, info

warden logs show a1b2c3d4 --report-on high

--min-confidence

string

Filter results to show only findings at or above confidence levelValues: off, high, medium, low

warden logs show a1b2c3d4 --min-confidence high

-v, --verbose

flag

Show detailed skill execution information

warden logs show a1b2c3d4 -v

--debug

boolean

default:"false"

Show debug information including token counts and latencies

warden logs show a1b2c3d4 --debug

Global Options

--quiet

boolean

default:"false"

Suppress non-error output

warden logs gc --quiet

--color / --no-color

boolean

Force color output on or off

warden logs --no-color

Log File Format

Log files are saved in JSONL (JSON Lines) format at .warden/logs/:

.warden/logs/
├── run-20260305-143022-a1b2c3d4.jsonl
├── run-20260305-150145-e5f6g7h8.jsonl
└── run-20260304-091233-i9j0k1l2.jsonl

Filename Format

run-<timestamp>-<short-run-id>.jsonl

timestamp: YYYYMMDD-HHMMSS in local time
short-run-id: First 8 characters of UUID

JSONL Structure

Each line is a JSON object. First line contains run metadata:

{
  "_meta": {
    "run": {
      "runId": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
      "traceId": "trace-xyz",
      "timestamp": "2026-03-05T14:30:22.000Z",
      "durationMs": 45123
    },
    "model": "claude-sonnet-4.5",
    "headSha": "abc1234def5678",
    "cwd": "/home/user/project"
  }
}

Subsequent lines contain skill reports:

{
  "skill": "security-review",
  "trigger": {"type": "pull_request"},
  "findings": [...],
  "durationMs": 12345,
  "usage": {
    "inputTokens": 1000,
    "outputTokens": 500,
    "cacheCreationTokens": 800,
    "cacheReadTokens": 200,
    "costUSD": 0.05
  }
}

Final line contains summary:

{
  "_summary": {
    "totalFindings": 4,
    "bySeverity": {"high": 2, "medium": 1, "low": 1},
    "run": {...},
    "usage": {...}
  }
}

Exit Codes

Success

Logs listed successfully
Results displayed successfully
Garbage collection completed

Error

Not in a git repository
Log file not found
Failed to parse log file
Invalid arguments

Examples

View Recent Runs

# List all runs
warden logs

# Show most recent run
warden logs show $(warden logs --json | jq -r '.[0].file')

Filter Results

# Show only high severity findings
warden logs show a1b2c3d4 --report-on high

# Show only high confidence findings
warden logs show a1b2c3d4 --min-confidence high

# Combine filters
warden logs show a1b2c3d4 --report-on high --min-confidence high

Export to JSON

# Export log list
warden logs --json > logs-summary.json

# Export specific run
warden logs show a1b2c3d4 --json > run-results.json

Compare Runs

# Show two runs side by side
warden logs show run1.jsonl > /tmp/run1.txt
warden logs show run2.jsonl > /tmp/run2.txt
diff /tmp/run1.txt /tmp/run2.txt

Clean Up Old Logs

# Remove logs older than retention period
warden logs gc

# See what will be removed (dry run)
find .warden/logs -name "*.jsonl" -mtime +30

CI/CD Integration

# Save run for archival
warden -o build-${CI_COMMIT_SHA}.jsonl

# Replay in later job
warden logs show build-${CI_COMMIT_SHA}.jsonl --json

Automatic Logging

Logs are automatically created when Warden runs successfully:

# This creates .warden/logs/run-<timestamp>-<id>.jsonl
warden

# Explicit output path (doesn't auto-save to .warden/logs/)
warden -o custom.jsonl

Disable auto-logging by setting in warden.toml:

[logs]
enabled = false

Log Retention

Configure retention in warden.toml:

[logs]
retentionDays = 30  # Default

Run warden logs gc periodically (e.g., in cron or CI) to clean up:

# In CI before analysis
warden logs gc
warden

GC is also run automatically before new analysis runs (non-blocking).

Use Cases

Debug Past Runs

# List recent runs
warden logs

# Replay with verbose output
warden logs show a1b2c3d4 -vv

# Generate shareable report
warden logs show a1b2c3d4 > findings.txt

# Or as JSON for processing
warden logs show a1b2c3d4 --json > findings.json

Compare Before/After

# Run before changes
warden
# Note the run ID

# Make changes
git commit -am "Fix security issues"

# Run again
warden

# Compare
warden logs show <before-id> --report-on high > before.txt
warden logs show <after-id> --report-on high > after.txt
diff before.txt after.txt

Cost Tracking

# View costs over time
warden logs --json | jq '.[] | {date: .timestamp, cost: .costUSD}'

# Total cost last 30 days
warden logs --json | jq '[.[] | .costUSD] | add'

warden - Run analysis (creates logs)
warden sync - Update remote skills

Library API

Types

CLI Commands

warden logs

warden logs

Usage

Subcommands

list (default)

show

gc

Options

For list

For show

Global Options

Log File Format

Filename Format

JSONL Structure

Exit Codes

Examples

View Recent Runs

Filter Results

Export to JSON

Compare Runs

Clean Up Old Logs

CI/CD Integration

Automatic Logging

Log Retention

Use Cases

Debug Past Runs

Compare Before/After

Cost Tracking

Build docs developers (and LLMs) love

Library API

Types

CLI Commands

​warden logs

​Usage

​Subcommands

​list (default)

​show

​gc

​Options

​For list

​For show

​Global Options

​Log File Format

​Filename Format

​JSONL Structure

​Exit Codes

​Examples

​View Recent Runs

​Filter Results

​Export to JSON

​Compare Runs

​Clean Up Old Logs

​CI/CD Integration

​Automatic Logging

​Log Retention

​Use Cases

​Debug Past Runs

​Share Results

​Compare Before/After

​Cost Tracking

​Related Commands

Build docs developers (and LLMs) love

warden logs

Usage

Subcommands

list (default)

show

gc

Options

For list

For show

Global Options

Log File Format

Filename Format

JSONL Structure

Exit Codes

Examples

View Recent Runs

Filter Results

Export to JSON

Compare Runs

Clean Up Old Logs

CI/CD Integration

Automatic Logging

Log Retention

Use Cases

Debug Past Runs

Share Results

Compare Before/After

Cost Tracking

Related Commands