@artemiskit/reports 0.2.2 → 0.2.4

package/CHANGELOG.md

@@ -1,5 +1,82 @@
 # @artemiskit/reports
 
+## 0.2.4
+
+### Patch Changes
+
+- 16604a6: ## New Features
+
+  ### Validate Command
+
+  New `artemiskit validate` command for validating scenario files without running them:
+
+  - **YAML syntax validation** - Catches formatting errors
+  - **Schema validation** - Validates against ArtemisKit schema using Zod
+  - **Semantic validation** - Detects duplicate case IDs, undefined variables
+  - **Warnings** - Identifies deprecated fields, missing descriptions, performance hints
+
+  Options:
+
+  - `--json` - Output results as JSON
+  - `--strict` - Treat warnings as errors
+  - `--quiet` - Only show errors
+  - `--export junit` - Export to JUnit XML for CI integration
+
+  ### JUnit XML Export
+
+  Added JUnit XML export support for CI/CD integration with Jenkins, GitHub Actions, GitLab CI, and other systems:
+
+  - `akit run scenarios/ --export junit` - Export run results
+  - `akit redteam scenarios/chatbot.yaml --export junit` - Export security test results
+  - `akit validate scenarios/ --export junit` - Export validation results
+
+  JUnit reports include:
+
+  - Test suite metadata (run ID, provider, model, success rate)
+  - Individual test cases with pass/fail status
+  - Failure details with matcher type and expected values
+  - Timing information for each test
+
+- Updated dependencies [16604a6]
+  - @artemiskit/core@0.2.4
+
+## 0.2.3
+
+### Patch Changes
+
+- 37403aa: ## v0.2.3 - Cost Tracking & Compliance Features
+
+  ### Cost Tracking
+
+  - **Automatic cost estimation**: Run results now include estimated API costs based on token usage and model pricing data
+  - **Cost display in output**: Summary output shows total tokens and estimated cost for each run
+  - **`--budget` flag**: Set a maximum budget in USD for `run`, `redteam`, and `stress` commands - the command fails (exit code 1) if the estimated cost exceeds the budget
+
+  ### History Enhancements
+
+  - **`--show-cost` flag**: Display cost column and total in `history` command output
+  - Cost data is stored with each run for historical tracking
+
+  ### Markdown Export
+
+  - **`--export markdown` flag**: Export run and redteam results to compliance-ready markdown format
+  - **`--export-output` flag**: Specify custom output directory for exports (default: `./artemis-exports`)
+  - Markdown reports include:
+    - Summary table with pass/fail rates, latency, token usage, and cost metrics
+    - Detailed results for failed test cases (run) or vulnerabilities found (redteam)
+    - Configuration used for the run
+    - Redaction summary (if enabled)
+    - Recommendations for remediation (redteam)
+
+  ### CI/CD Integration
+
+  - Budget enforcement in pipelines: `akit run scenarios/ --ci --budget 5.00`
+  - Cost tracking in CI summary output with `ARTEMISKIT_COST_USD` variable
+  - Automatic markdown report generation for compliance documentation
+
+- Updated dependencies [37403aa]
+  - @artemiskit/core@0.2.3
+
 ## 0.2.2
 
 ### Patch Changes

package/dist/index.d.ts

@@ -7,4 +7,6 @@ export { generateJSONReport, type JSONReportOptions } from './json/generator';
 export { generateRedTeamHTMLReport } from './html/redteam-generator';
 export { generateStressHTMLReport } from './html/stress-generator';
 export { generateCompareHTMLReport, buildComparisonData, type ComparisonData, type CaseComparison, } from './html/compare-generator';
+export { generateMarkdownReport, generateRedTeamMarkdownReport, type MarkdownReportOptions, } from './markdown/generator';
+export { generateJUnitReport, generateRedTeamJUnitReport, generateValidationJUnitReport, type JUnitReportOptions, } from './junit/generator';
 //# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAC;AACtD,OAAO,EAAE,kBAAkB,EAAE,KAAK,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAG9E,OAAO,EAAE,yBAAyB,EAAE,MAAM,0BAA0B,CAAC;AAGrE,OAAO,EAAE,wBAAwB,EAAE,MAAM,yBAAyB,CAAC;AAGnE,OAAO,EACL,yBAAyB,EACzB,mBAAmB,EACnB,KAAK,cAAc,EACnB,KAAK,cAAc,GACpB,MAAM,0BAA0B,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAC;AACtD,OAAO,EAAE,kBAAkB,EAAE,KAAK,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAG9E,OAAO,EAAE,yBAAyB,EAAE,MAAM,0BAA0B,CAAC;AAGrE,OAAO,EAAE,wBAAwB,EAAE,MAAM,yBAAyB,CAAC;AAGnE,OAAO,EACL,yBAAyB,EACzB,mBAAmB,EACnB,KAAK,cAAc,EACnB,KAAK,cAAc,GACpB,MAAM,0BAA0B,CAAC;AAGlC,OAAO,EACL,sBAAsB,EACtB,6BAA6B,EAC7B,KAAK,qBAAqB,GAC3B,MAAM,sBAAsB,CAAC;AAG9B,OAAO,EACL,mBAAmB,EACnB,0BAA0B,EAC1B,6BAA6B,EAC7B,KAAK,kBAAkB,GACxB,MAAM,mBAAmB,CAAC"}

package/dist/index.js

@@ -7901,9 +7901,466 @@ function generateCompareHTMLReport(baseline, current) {
   const template = import_handlebars4.default.compile(COMPARE_HTML_TEMPLATE);
   return template({ data });
 }
+// src/markdown/generator.ts
+function truncate(text, maxLength) {
+  if (text.length <= maxLength)
+    return text;
+  return `${text.slice(0, maxLength)}...`;
+}
+function formatCostMd(costUsd) {
+  if (costUsd < 0.01) {
+    return `$${(costUsd * 100).toFixed(4)} cents`;
+  }
+  if (costUsd < 1) {
+    return `$${costUsd.toFixed(4)}`;
+  }
+  return `$${costUsd.toFixed(2)}`;
+}
+function formatDuration(ms) {
+  if (ms < 1000)
+    return `${ms}ms`;
+  if (ms < 60000)
+    return `${(ms / 1000).toFixed(1)}s`;
+  const minutes = Math.floor(ms / 60000);
+  const seconds = (ms % 60000 / 1000).toFixed(0);
+  return `${minutes}m ${seconds}s`;
+}
+function generateMarkdownReport(manifest, options = {}) {
+  const { includeDetails = true, truncateAt = 500 } = options;
+  const lines = [];
+  lines.push("# ArtemisKit Test Results");
+  lines.push("");
+  lines.push(`**Scenario:** ${manifest.config.scenario}`);
+  lines.push(`**Run ID:** ${manifest.run_id}`);
+  lines.push(`**Date:** ${new Date(manifest.start_time).toISOString()}`);
+  lines.push(`**Provider:** ${manifest.config.provider}${manifest.config.model ? ` (${manifest.config.model})` : ""}`);
+  lines.push("");
+  lines.push("---");
+  lines.push("");
+  lines.push("## Summary");
+  lines.push("");
+  lines.push("| Metric | Value |");
+  lines.push("|--------|-------|");
+  lines.push(`| Total Cases | ${manifest.metrics.total_cases} |`);
+  lines.push(`| Passed | ${manifest.metrics.passed_cases} (${(manifest.metrics.success_rate * 100).toFixed(1)}%) |`);
+  lines.push(`| Failed | ${manifest.metrics.failed_cases} |`);
+  lines.push(`| Duration | ${formatDuration(manifest.duration_ms)} |`);
+  lines.push(`| Median Latency | ${manifest.metrics.median_latency_ms}ms |`);
+  lines.push(`| P95 Latency | ${manifest.metrics.p95_latency_ms}ms |`);
+  lines.push(`| Total Tokens | ${manifest.metrics.total_tokens.toLocaleString()} |`);
+  if (manifest.metrics.cost) {
+    lines.push(`| Estimated Cost | ${formatCostMd(manifest.metrics.cost.total_usd)} |`);
+  }
+  lines.push("");
+  lines.push("---");
+  lines.push("");
+  lines.push("## Results by Case");
+  lines.push("");
+  const passed = manifest.cases.filter((c) => c.ok);
+  lines.push(`### Passed (${passed.length})`);
+  lines.push("");
+  if (passed.length > 0) {
+    lines.push("<details>");
+    lines.push("<summary>Click to expand passed cases</summary>");
+    lines.push("");
+    lines.push("| Case ID | Latency | Tokens | Score |");
+    lines.push("|---------|---------|--------|-------|");
+    for (const c of passed) {
+      lines.push(`| ${c.id} | ${formatDuration(c.latencyMs)} | ${c.tokens?.total || "-"} | ${(c.score * 100).toFixed(0)}% |`);
+    }
+    lines.push("");
+    lines.push("</details>");
+  } else {
+    lines.push("_No passed cases_");
+  }
+  lines.push("");
+  const failed = manifest.cases.filter((c) => !c.ok);
+  lines.push(`### Failed (${failed.length})`);
+  lines.push("");
+  if (failed.length > 0) {
+    for (const c of failed) {
+      lines.push(`#### \`${c.id}\``);
+      lines.push("");
+      if (includeDetails) {
+        const promptStr = typeof c.prompt === "string" ? c.prompt : JSON.stringify(c.prompt, null, 2);
+        lines.push("**Prompt:**");
+        lines.push("```");
+        lines.push(truncate(promptStr, truncateAt));
+        lines.push("```");
+        lines.push("");
+        lines.push("**Expected:**");
+        lines.push(`- Type: \`${c.matcherType}\``);
+        lines.push("```json");
+        lines.push(truncate(JSON.stringify(c.expected, null, 2), truncateAt));
+        lines.push("```");
+        lines.push("");
+        lines.push("**Actual Response:**");
+        lines.push("```");
+        lines.push(truncate(c.response || "(empty)", truncateAt));
+        lines.push("```");
+        lines.push("");
+      }
+      lines.push(`**Reason:** ${c.reason || "Unknown"}`);
+      lines.push("");
+      lines.push("---");
+      lines.push("");
+    }
+  } else {
+    lines.push("_No failed cases_");
+    lines.push("");
+  }
+  if (manifest.resolved_config) {
+    lines.push("## Configuration");
+    lines.push("");
+    lines.push("```yaml");
+    lines.push(`provider: ${manifest.resolved_config.provider}`);
+    if (manifest.resolved_config.model) {
+      lines.push(`model: ${manifest.resolved_config.model}`);
+    }
+    if (manifest.resolved_config.temperature !== undefined) {
+      lines.push(`temperature: ${manifest.resolved_config.temperature}`);
+    }
+    if (manifest.resolved_config.max_tokens !== undefined) {
+      lines.push(`max_tokens: ${manifest.resolved_config.max_tokens}`);
+    }
+    lines.push("```");
+    lines.push("");
+  }
+  if (manifest.redaction?.enabled) {
+    lines.push("## Redaction");
+    lines.push("");
+    lines.push(`- **Patterns Used:** ${manifest.redaction.patternsUsed.join(", ")}`);
+    lines.push(`- **Prompts Redacted:** ${manifest.redaction.summary.promptsRedacted}`);
+    lines.push(`- **Responses Redacted:** ${manifest.redaction.summary.responsesRedacted}`);
+    lines.push(`- **Total Redactions:** ${manifest.redaction.summary.totalRedactions}`);
+    lines.push("");
+  }
+  lines.push("---");
+  lines.push("");
+  lines.push(`*Generated by [ArtemisKit](https://artemiskit.vercel.app) v${manifest.version}*`);
+  return lines.join(`
+`);
+}
+function generateRedTeamMarkdownReport(manifest, options = {}) {
+  const { includeDetails = true, truncateAt = 500 } = options;
+  const lines = [];
+  lines.push("# ArtemisKit Security Report");
+  lines.push("");
+  lines.push(`**Scenario:** ${manifest.config.scenario}`);
+  lines.push(`**Run ID:** ${manifest.run_id}`);
+  lines.push(`**Date:** ${new Date(manifest.start_time).toISOString()}`);
+  lines.push("**Test Type:** Red Team Security Scan");
+  lines.push("");
+  lines.push("---");
+  lines.push("");
+  const testable = manifest.metrics.total_tests - manifest.metrics.error_responses;
+  const defenseRate = manifest.metrics.defense_rate * 100;
+  let riskLevel;
+  let riskEmoji;
+  if (defenseRate >= 95) {
+    riskLevel = "LOW";
+    riskEmoji = "";
+  } else if (defenseRate >= 80) {
+    riskLevel = "MEDIUM";
+    riskEmoji = "";
+  } else if (defenseRate >= 50) {
+    riskLevel = "HIGH";
+    riskEmoji = "";
+  } else {
+    riskLevel = "CRITICAL";
+    riskEmoji = "";
+  }
+  lines.push("## Security Summary");
+  lines.push("");
+  lines.push("| Metric | Value |");
+  lines.push("|--------|-------|");
+  lines.push(`| Defense Rate | ${defenseRate.toFixed(1)}% |`);
+  lines.push(`| Total Attacks | ${manifest.metrics.total_tests} |`);
+  lines.push(`| Defended | ${manifest.metrics.defended} |`);
+  lines.push(`| Safe Responses | ${manifest.metrics.safe_responses} |`);
+  lines.push(`| Blocked | ${manifest.metrics.blocked_responses} |`);
+  lines.push(`| Vulnerabilities | ${manifest.metrics.unsafe_responses} |`);
+  lines.push(`| Errors | ${manifest.metrics.error_responses} |`);
+  lines.push(`| Risk Level | **${riskEmoji} ${riskLevel}** |`);
+  lines.push("");
+  lines.push("---");
+  lines.push("");
+  const { by_severity } = manifest.metrics;
+  if (manifest.metrics.unsafe_responses > 0) {
+    lines.push("## Severity Breakdown");
+    lines.push("");
+    lines.push("| Severity | Count | Description |");
+    lines.push("|----------|-------|-------------|");
+    if (by_severity.critical > 0) {
+      lines.push(`| Critical | ${by_severity.critical} | Severe security vulnerability |`);
+    }
+    if (by_severity.high > 0) {
+      lines.push(`| High | ${by_severity.high} | Significant security concern |`);
+    }
+    if (by_severity.medium > 0) {
+      lines.push(`| Medium | ${by_severity.medium} | Moderate security issue |`);
+    }
+    if (by_severity.low > 0) {
+      lines.push(`| Low | ${by_severity.low} | Minor security concern |`);
+    }
+    lines.push("");
+    lines.push("---");
+    lines.push("");
+  }
+  const unsafeResults = manifest.results.filter((r) => r.status === "unsafe");
+  if (unsafeResults.length > 0) {
+    lines.push("## Vulnerabilities Found");
+    lines.push("");
+    for (let i = 0;i < unsafeResults.length; i++) {
+      const result = unsafeResults[i];
+      lines.push(`### ${i + 1}. ${result.mutation} (${result.severity.toUpperCase()})`);
+      lines.push("");
+      lines.push(`**Case ID:** ${result.caseId}`);
+      lines.push(`**Severity:** ${result.severity.toUpperCase()}`);
+      lines.push(`**Mutation:** ${result.mutation}`);
+      lines.push("");
+      if (includeDetails) {
+        lines.push("**Attack Prompt:**");
+        lines.push("```");
+        lines.push(truncate(result.prompt, truncateAt));
+        lines.push("```");
+        lines.push("");
+        lines.push("**Model Response:**");
+        lines.push("```");
+        lines.push(truncate(result.response || "(empty)", truncateAt));
+        lines.push("```");
+        lines.push("");
+      }
+      if (result.reasons.length > 0) {
+        lines.push("**Reasons:**");
+        for (const reason of result.reasons) {
+          lines.push(`- ${reason}`);
+        }
+        lines.push("");
+      }
+      lines.push("---");
+      lines.push("");
+    }
+  }
+  lines.push("## Recommendations");
+  lines.push("");
+  if (manifest.metrics.unsafe_responses === 0) {
+    lines.push("- All attacks were successfully defended");
+    lines.push("- Continue monitoring and testing regularly");
+  } else {
+    if (by_severity.critical > 0 || by_severity.high > 0) {
+      lines.push("- **High Priority:** Review and address critical/high severity vulnerabilities immediately");
+    }
+    if (manifest.metrics.unsafe_responses > testable * 0.1) {
+      lines.push("- Consider implementing additional input validation and output filtering");
+    }
+    lines.push("- Review system prompts for potential information leakage");
+    lines.push("- Implement role-play and persona detection");
+    lines.push("- Add output filtering for dangerous content patterns");
+  }
+  lines.push("");
+  lines.push("---");
+  lines.push("");
+  lines.push(`*Generated by [ArtemisKit](https://artemiskit.vercel.app) v${manifest.version}*`);
+  return lines.join(`
+`);
+}
+// src/junit/generator.ts
+function escapeXml(str) {
+  const invalidXmlChars = /[\x00-\x08\x0B\x0C\x0E-\x1F]/g;
+  return str.replace(/&/g, "&amp;").replace(/</g, "&lt;").replace(/>/g, "&gt;").replace(/"/g, "&quot;").replace(/'/g, "&apos;").replace(invalidXmlChars, "");
+}
+function truncate2(text, maxLength) {
+  if (text.length <= maxLength)
+    return text;
+  return `${text.slice(0, maxLength)}...(truncated)`;
+}
+function formatTimestamp(dateStr) {
+  return new Date(dateStr).toISOString();
+}
+function generateJUnitReport(manifest, options = {}) {
+  const {
+    suiteName = manifest.config.scenario,
+    includeSystemOut = true,
+    includeSystemErr = true,
+    maxOutputLength = 2000
+  } = options;
+  const lines = [];
+  lines.push('<?xml version="1.0" encoding="UTF-8"?>');
+  const tests = manifest.metrics.total_cases;
+  const failures = manifest.metrics.failed_cases;
+  const errors = 0;
+  const skipped = 0;
+  const time = manifest.duration_ms / 1000;
+  lines.push(`<testsuite name="${escapeXml(suiteName)}" ` + `tests="${tests}" failures="${failures}" errors="${errors}" skipped="${skipped}" ` + `time="${time.toFixed(3)}" timestamp="${formatTimestamp(manifest.start_time)}">`);
+  lines.push("  <properties>");
+  lines.push(`    <property name="artemis.run_id" value="${escapeXml(manifest.run_id)}" />`);
+  lines.push(`    <property name="artemis.version" value="${escapeXml(manifest.version)}" />`);
+  lines.push(`    <property name="artemis.provider" value="${escapeXml(manifest.config.provider)}" />`);
+  if (manifest.config.model) {
+    lines.push(`    <property name="artemis.model" value="${escapeXml(manifest.config.model)}" />`);
+  }
+  lines.push(`    <property name="artemis.success_rate" value="${(manifest.metrics.success_rate * 100).toFixed(1)}%" />`);
+  lines.push(`    <property name="artemis.total_tokens" value="${manifest.metrics.total_tokens}" />`);
+  if (manifest.metrics.cost) {
+    lines.push(`    <property name="artemis.cost_usd" value="${manifest.metrics.cost.total_usd.toFixed(6)}" />`);
+  }
+  lines.push("  </properties>");
+  for (const testCase of manifest.cases) {
+    const className = escapeXml(suiteName);
+    const testName = escapeXml(testCase.id);
+    const testTime = testCase.latencyMs / 1000;
+    lines.push(`  <testcase classname="${className}" name="${testName}" time="${testTime.toFixed(3)}">`);
+    if (!testCase.ok) {
+      const failureMessage = escapeXml(testCase.reason || "Test failed");
+      const failureType = escapeXml(testCase.matcherType);
+      lines.push(`    <failure message="${failureMessage}" type="${failureType}">`);
+      const details = [];
+      details.push(`Matcher Type: ${testCase.matcherType}`);
+      details.push(`Expected: ${JSON.stringify(testCase.expected, null, 2)}`);
+      details.push(`Score: ${(testCase.score * 100).toFixed(1)}%`);
+      if (testCase.reason) {
+        details.push(`Reason: ${testCase.reason}`);
+      }
+      lines.push(escapeXml(details.join(`
+`)));
+      lines.push("    </failure>");
+    }
+    if (includeSystemOut && testCase.response) {
+      lines.push("    <system-out>");
+      lines.push(`<![CDATA[${truncate2(testCase.response, maxOutputLength)}]]>`);
+      lines.push("    </system-out>");
+    }
+    if (includeSystemErr && !testCase.ok && testCase.reason) {
+      lines.push("    <system-err>");
+      const errorDetails = [];
+      errorDetails.push(`Error: ${testCase.reason}`);
+      const promptStr = typeof testCase.prompt === "string" ? testCase.prompt : JSON.stringify(testCase.prompt);
+      errorDetails.push(`Prompt: ${truncate2(promptStr, maxOutputLength / 2)}`);
+      lines.push(`<![CDATA[${errorDetails.join(`
+`)}]]>`);
+      lines.push("    </system-err>");
+    }
+    lines.push("  </testcase>");
+  }
+  lines.push("</testsuite>");
+  return lines.join(`
+`);
+}
+function generateRedTeamJUnitReport(manifest, options = {}) {
+  const {
+    suiteName = `RedTeam: ${manifest.config.scenario}`,
+    includeSystemOut = true,
+    includeSystemErr = true,
+    maxOutputLength = 2000
+  } = options;
+  const lines = [];
+  lines.push('<?xml version="1.0" encoding="UTF-8"?>');
+  const tests = manifest.metrics.total_tests;
+  const failures = manifest.metrics.unsafe_responses;
+  const errors = manifest.metrics.error_responses;
+  const skipped = 0;
+  const time = manifest.duration_ms / 1000;
+  lines.push(`<testsuite name="${escapeXml(suiteName)}" ` + `tests="${tests}" failures="${failures}" errors="${errors}" skipped="${skipped}" ` + `time="${time.toFixed(3)}" timestamp="${formatTimestamp(manifest.start_time)}">`);
+  lines.push("  <properties>");
+  lines.push(`    <property name="artemis.run_id" value="${escapeXml(manifest.run_id)}" />`);
+  lines.push(`    <property name="artemis.version" value="${escapeXml(manifest.version)}" />`);
+  lines.push(`    <property name="artemis.test_type" value="redteam" />`);
+  lines.push(`    <property name="artemis.defense_rate" value="${(manifest.metrics.defense_rate * 100).toFixed(1)}%" />`);
+  lines.push(`    <property name="artemis.safe_responses" value="${manifest.metrics.safe_responses}" />`);
+  lines.push(`    <property name="artemis.blocked_responses" value="${manifest.metrics.blocked_responses}" />`);
+  lines.push(`    <property name="artemis.unsafe_responses" value="${manifest.metrics.unsafe_responses}" />`);
+  lines.push("  </properties>");
+  for (const result of manifest.results) {
+    const className = escapeXml(suiteName);
+    const testName = escapeXml(`${result.caseId} (${result.mutation})`);
+    const testTime = (result.latencyMs || 0) / 1000;
+    lines.push(`  <testcase classname="${className}" name="${testName}" time="${testTime.toFixed(3)}">`);
+    if (result.status === "unsafe") {
+      const failureMessage = escapeXml(`Security vulnerability: ${result.mutation}`);
+      const failureType = escapeXml(`severity:${result.severity}`);
+      lines.push(`    <failure message="${failureMessage}" type="${failureType}">`);
+      const details = [];
+      details.push(`Severity: ${result.severity.toUpperCase()}`);
+      details.push(`Mutation: ${result.mutation}`);
+      if (result.reasons.length > 0) {
+        details.push(`Reasons: ${result.reasons.join(", ")}`);
+      }
+      lines.push(escapeXml(details.join(`
+`)));
+      lines.push("    </failure>");
+    } else if (result.status === "error") {
+      lines.push(`    <error message="${escapeXml(result.response || "Error during test")}" type="error">`);
+      lines.push(escapeXml(`Attack: ${result.mutation}
+Case: ${result.caseId}`));
+      lines.push("    </error>");
+    }
+    if (includeSystemOut && result.response) {
+      lines.push("    <system-out>");
+      lines.push(`<![CDATA[${truncate2(result.response, maxOutputLength)}]]>`);
+      lines.push("    </system-out>");
+    }
+    if (includeSystemErr && result.status === "unsafe") {
+      lines.push("    <system-err>");
+      const errDetails = [];
+      errDetails.push(`Attack Prompt: ${truncate2(result.prompt, maxOutputLength / 2)}`);
+      errDetails.push(`Severity: ${result.severity.toUpperCase()}`);
+      lines.push(`<![CDATA[${errDetails.join(`
+`)}]]>`);
+      lines.push("    </system-err>");
+    }
+    lines.push("  </testcase>");
+  }
+  lines.push("</testsuite>");
+  return lines.join(`
+`);
+}
+function generateValidationJUnitReport(results, options = {}) {
+  const { suiteName = "ArtemisKit Validation" } = options;
+  const lines = [];
+  lines.push('<?xml version="1.0" encoding="UTF-8"?>');
+  const tests = results.length;
+  const failures = results.filter((r) => !r.valid).length;
+  const errors = 0;
+  const skipped = 0;
+  lines.push(`<testsuite name="${escapeXml(suiteName)}" ` + `tests="${tests}" failures="${failures}" errors="${errors}" skipped="${skipped}" ` + `time="0" timestamp="${new Date().toISOString()}">`);
+  for (const result of results) {
+    const className = escapeXml(suiteName);
+    const testName = escapeXml(result.file);
+    lines.push(`  <testcase classname="${className}" name="${testName}" time="0">`);
+    if (!result.valid) {
+      const errorMessages = result.errors.map((e) => `Line ${e.line}: ${e.message}`).join("; ");
+      lines.push(`    <failure message="${escapeXml(errorMessages)}" type="validation">`);
+      const details = [];
+      for (const error of result.errors) {
+        details.push(`[${error.rule}] Line ${error.line}: ${error.message}`);
+      }
+      lines.push(escapeXml(details.join(`
+`)));
+      lines.push("    </failure>");
+    }
+    if (result.warnings.length > 0) {
+      lines.push("    <system-err>");
+      const warningDetails = result.warnings.map((w) => `[${w.rule}] Line ${w.line}: ${w.message}`).join(`
+`);
+      lines.push(`<![CDATA[Warnings:
+${warningDetails}]]>`);
+      lines.push("    </system-err>");
+    }
+    lines.push("  </testcase>");
+  }
+  lines.push("</testsuite>");
+  return lines.join(`
+`);
+}
 export {
+  generateValidationJUnitReport,
   generateStressHTMLReport,
+  generateRedTeamMarkdownReport,
+  generateRedTeamJUnitReport,
   generateRedTeamHTMLReport,
+  generateMarkdownReport,
+  generateJUnitReport,
   generateJSONReport,
   generateHTMLReport,
   generateCompareHTMLReport,

package/dist/junit/generator.d.ts

@@ -0,0 +1,44 @@
+/**
+ * JUnit XML Report Generator
+ *
+ * Generates JUnit-compatible XML reports for CI/CD integration.
+ * Follows the JUnit XML format specification for compatibility with
+ * Jenkins, GitHub Actions, GitLab CI, and other CI systems.
+ */
+import type { RedTeamManifest, RunManifest } from '@artemiskit/core';
+export interface JUnitReportOptions {
+    /** Test suite name (defaults to scenario name) */
+    suiteName?: string;
+    /** Include system-out with response content */
+    includeSystemOut?: boolean;
+    /** Include system-err with error details */
+    includeSystemErr?: boolean;
+    /** Maximum content length for outputs */
+    maxOutputLength?: number;
+}
+/**
+ * Generate JUnit XML report for a standard run
+ */
+export declare function generateJUnitReport(manifest: RunManifest, options?: JUnitReportOptions): string;
+/**
+ * Generate JUnit XML report for red team results
+ */
+export declare function generateRedTeamJUnitReport(manifest: RedTeamManifest, options?: JUnitReportOptions): string;
+/**
+ * Generate JUnit XML for validation results
+ */
+export declare function generateValidationJUnitReport(results: Array<{
+    file: string;
+    valid: boolean;
+    errors: Array<{
+        line: number;
+        message: string;
+        rule: string;
+    }>;
+    warnings: Array<{
+        line: number;
+        message: string;
+        rule: string;
+    }>;
+}>, options?: JUnitReportOptions): string;
+//# sourceMappingURL=generator.d.ts.map

package/dist/junit/generator.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"generator.d.ts","sourceRoot":"","sources":["../../src/junit/generator.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EAAE,eAAe,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAC;AAErE,MAAM,WAAW,kBAAkB;IACjC,kDAAkD;IAClD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,+CAA+C;IAC/C,gBAAgB,CAAC,EAAE,OAAO,CAAC;IAC3B,4CAA4C;IAC5C,gBAAgB,CAAC,EAAE,OAAO,CAAC;IAC3B,yCAAyC;IACzC,eAAe,CAAC,EAAE,MAAM,CAAC;CAC1B;AAmCD;;GAEG;AACH,wBAAgB,mBAAmB,CACjC,QAAQ,EAAE,WAAW,EACrB,OAAO,GAAE,kBAAuB,GAC/B,MAAM,CA0GR;AAED;;GAEG;AACH,wBAAgB,0BAA0B,CACxC,QAAQ,EAAE,eAAe,EACzB,OAAO,GAAE,kBAAuB,GAC/B,MAAM,CAyGR;AAED;;GAEG;AACH,wBAAgB,6BAA6B,CAC3C,OAAO,EAAE,KAAK,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;IACb,KAAK,EAAE,OAAO,CAAC;IACf,MAAM,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC/D,QAAQ,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;CAClE,CAAC,EACF,OAAO,GAAE,kBAAuB,GAC/B,MAAM,CAyDR"}

package/dist/markdown/generator.d.ts

@@ -0,0 +1,21 @@
+/**
+ * Markdown Report Generator
+ *
+ * Generates documentation-friendly markdown reports for compliance and audit trails.
+ */
+import type { RedTeamManifest, RunManifest } from '@artemiskit/core';
+export interface MarkdownReportOptions {
+    /** Include full prompt/response details for failed cases */
+    includeDetails?: boolean;
+    /** Maximum characters to show for prompts/responses */
+    truncateAt?: number;
+}
+/**
+ * Generate markdown report for a standard run
+ */
+export declare function generateMarkdownReport(manifest: RunManifest, options?: MarkdownReportOptions): string;
+/**
+ * Generate markdown report for red team results
+ */
+export declare function generateRedTeamMarkdownReport(manifest: RedTeamManifest, options?: MarkdownReportOptions): string;
+//# sourceMappingURL=generator.d.ts.map

package/dist/markdown/generator.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"generator.d.ts","sourceRoot":"","sources":["../../src/markdown/generator.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,KAAK,EAAE,eAAe,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAC;AAErE,MAAM,WAAW,qBAAqB;IACpC,4DAA4D;IAC5D,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB,uDAAuD;IACvD,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAkCD;;GAEG;AACH,wBAAgB,sBAAsB,CACpC,QAAQ,EAAE,WAAW,EACrB,OAAO,GAAE,qBAA0B,GAClC,MAAM,CAuJR;AAED;;GAEG;AACH,wBAAgB,6BAA6B,CAC3C,QAAQ,EAAE,eAAe,EACzB,OAAO,GAAE,qBAA0B,GAClC,MAAM,CAmJR"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@artemiskit/reports",
-  "version": "0.2.2",
+  "version": "0.2.4",
   "description": "HTML report generation for ArtemisKit LLM evaluation toolkit",
   "type": "module",
   "license": "Apache-2.0",
@@ -30,7 +30,7 @@
     "test": "bun test"
   },
   "dependencies": {
-    "@artemiskit/core": "workspace:*",
+    "@artemiskit/core": "0.2.4",
     "handlebars": "^4.7.8"
   },
   "devDependencies": {

package/src/index.ts

@@ -20,3 +20,18 @@ export {
   type ComparisonData,
   type CaseComparison,
 } from './html/compare-generator';
+
+// Markdown reports
+export {
+  generateMarkdownReport,
+  generateRedTeamMarkdownReport,
+  type MarkdownReportOptions,
+} from './markdown/generator';
+
+// JUnit XML reports (CI integration)
+export {
+  generateJUnitReport,
+  generateRedTeamJUnitReport,
+  generateValidationJUnitReport,
+  type JUnitReportOptions,
+} from './junit/generator';