Output Formats

skill-issue supports three output formats, each suited for different use cases.

Table (Default)

Human-readable formatted table with colored severity indicators. Best for interactive use.

skill-issue /path/to/skill

╭──────────┬────────────┬──────────────┬──────┬──────────────────────────────────────╮
│ Severity │ Rule       │ File         │ Line │ Message                              │
├──────────┼────────────┼──────────────┼──────┼──────────────────────────────────────┤
│ ERROR    │ SL-SEC-001 │ README.md    │ 24   │ Possible API key detected            │
│ WARNING  │ SL-HID-002 │ README.md    │ 12   │ HTML comment may hide instructions   │
│ INFO     │ SL-NET-001 │ config.yml   │ 8    │ URL found in skill content           │
╰──────────┴────────────┴──────────────┴──────┴──────────────────────────────────────╯

Found 3 issue(s): 1 error(s), 1 warning(s), 1 info(s)

Severity indicators are color-coded:

• ERROR — Red
• WARNING — Yellow
• INFO — Blue

Verbose Mode

Add --verbose for detailed rule information:

skill-issue --verbose /path/to/skill

This shows the full matched text and rule description for each finding.

JSON

Structured JSON output for programmatic parsing. Best for CI/CD pipelines and custom tooling.

skill-issue --format json /path/to/skill

{
  "version": "0.1.0",
  "skill_path": "/path/to/skill",
  "findings": [
    {
      "rule_id": "SL-SEC-001",
      "rule_name": "API Key Pattern",
      "severity": "error",
      "message": "Possible API key detected: api_key = \"sk-1234567890abcdefghijklmnop\"",
      "location": {
        "file": "README.md",
        "line": 24,
        "column": 1
      },
      "matched_text": "api_key = \"sk-1234567890abcdefghijklmnop\""
    }
  ],
  "summary": {
    "total": 1,
    "errors": 1,
    "warnings": 0,
    "info": 0
  }
}

JSON Schema

Field	Type	Description
version	string	skill-issue version
skill_path	string	Absolute path to scanned directory
findings	array	List of security findings
findings[].rule_id	string	Unique rule identifier
findings[].rule_name	string	Human-readable rule name
findings[].severity	string	error, warning, or info
findings[].message	string	Detailed finding message
findings[].location.file	string	Relative file path
findings[].location.line	number	Line number (1-indexed)
findings[].location.column	number	Column number (1-indexed)
findings[].matched_text	string	Text that triggered the rule
summary.total	number	Total findings count
summary.errors	number	Error count
summary.warnings	number	Warning count
summary.info	number	Info count

Parsing with jq

# Count errors
skill-issue --format json | jq '.summary.errors'

# List all error-level findings
skill-issue --format json | jq '.findings[] | select(.severity == "error")'

# Get unique rule IDs that fired
skill-issue --format json | jq '[.findings[].rule_id] | unique'

# Extract file:line pairs
skill-issue --format json | jq -r '.findings[] | "\(.location.file):\(.location.line)"'

SARIF

Standards-based SARIF 2.1.0 format for integration with GitHub Code Scanning, Azure DevOps, and other security tools.

skill-issue --format sarif /path/to/skill > results.sarif

{
  "$schema": "https://json.schemastore.org/sarif-2.1.0.json",
  "version": "2.1.0",
  "runs": [
    {
      "tool": {
        "driver": {
          "name": "skill-issue",
          "version": "0.1.0",
          "rules": [...]
        }
      },
      "results": [...]
    }
  ]
}

GitHub Code Scanning

Upload SARIF results to GitHub for inline PR annotations:

- name: Upload SARIF
  uses: github/codeql-action/upload-sarif@v3
  with:
    sarif_file: results.sarif

Findings appear in the Security tab and as inline comments on pull requests.

Comparing Formats

Feature	Table	JSON	SARIF
Human-readable	Yes	No	No
Parseable	No	Yes	Yes
Color-coded	Yes	No	No
GitHub integration	No	No	Yes
CI/CD friendly	No	Yes	Yes
Verbose mode	Yes	N/A	N/A

Quiet Mode

Combine --quiet with any format to suppress non-finding output:

# Only show findings, no header or summary
skill-issue --quiet --format table

# Clean JSON with no extra output
skill-issue --quiet --format json > results.json