mirror of
https://github.com/KeygraphHQ/shannon.git
synced 2026-10-03 06:46:49 +02:00
* feat(worker): migrate agent runtime from Claude Agent SDK to pi harness * feat: remove Google Vertex AI provider support * fix(worker): route Bedrock and custom-base-URL providers from env * feat(prompts): instruct agents to call submit_exploitation_queue and submit_auth_result * fix(worker): count sub-agent cost and surface compaction failures * refactor(worker): rename claude-executor to pi-executor * feat(worker): pi-event-driven output formatting * fix(worker): gate adaptive thinking to Opus models, drop CLAUDE_THINKING_LEVEL * fix(worker): restore minLength/minItems on vuln-collector schemas * feat(worker): give task sub-agent write+bash, align tool descriptions * feat(worker): add glob custom tool and route code_path globs to it * refactor(prompts): use pi tool names (task, todo_write, read, bash, glob) * refactor(prompts): drop stale MCP terminology for collector tools * refactor(prompts): drop collector server names from deliverable instructions * fix(worker): restore minLength/minItems on pre-recon and exploit collector schemas * feat(worker): load playwright-cli skill via pi resource loader * refactor(cli): remove CLAUDE_CODE_MAX_OUTPUT_TOKENS config * build: drop @anthropic-ai/claude-code from worker image * docs: remove vertex references from llms context * docs(worker): update stale sdk comments * refactor(worker): unify provider precedence between preflight and executor * feat(worker): enforce bounded bash timeouts via pi extension * ci: bump the beta release line to 2.0.0 (#356) * fix(cli): pin npx command hints to beta tag * fix: render agent deliverables before the success commit so resume preserves them (#377) * feat(cli): restructure run folder and improve terminal UX (#383) * feat: surface report at run root and nest run internals under .shannon * feat: use plain-language wording in user-facing terminal messages * feat(cli): guide users to watch scan progress and surface report path on start * docs: sync run-folder layout and CLI wording across docs and comments * feat(cli): add version command reporting package version or git SHA * feat(cli): detect TTY for interactive prompts, color, and progress output * docs: document --yes flag, version command, and tty module * fix(cli): FORCE_COLOR precedence and plain uninstall --yes output * fix(cli): respect empty NO_COLOR * fix(cli): let NO_COLOR take precedence over FORCE_COLOR * docs: mark claude-code-router integration as removed * refactor(worker): converge shared core with shannon-oss (#388) * fix(worker): port keygraph shared-core correctness fixes * refactor(worker): adopt collectors/ and ai/pi/ layout; add task budget cap and cancellation * refactor(worker): drop inconsistent Collector "Server" suffix * refactor(worker): drop unused providerConfig/apiKey seams, resolve credentials from env only * refactor(worker): port oss code_path pattern expansion + external_directory allow * fix(worker): preserve dotfile paths in code_path avoid patterns (.env no longer stripped to env) * feat(worker): render Unprocessed Vulnerabilities section in exploit deliverable (align with oss) * feat(worker): request set_blind_spots for all vuln classes (align auth/ssrf with production prompts) * refactor(worker): adopt unified permissionSystem* naming and helper layout * refactor(worker): inline blind_spots into vuln deliverable section array * chore(worker): drop unused zod dependency (tree is typebox-native) * fix(worker): normalize base32 TOTP secret to accept padding and whitespace * refactor(worker): adopt shared toolResult helper and flatSchema naming in collectors * refactor(worker): use undefined over null in queue-schema builders * docs(worker): converge renderer/collector doc comments to current pi terminology * refactor(worker): adopt schema.ts cleanInput/stringEnum helpers in collectors * feat(worker): converge exploit-collector/renderer with vendored; capture and render overview for blocked findings * refactor(worker): converge session-tools/pipeline/exploitation-checker with vendored * refactor(worker): converge task-tool usage reporting with vendored onUsage callback * refactor(worker): converge structured output onto a submitTool executor channel * docs(worker): expand exploit-renderer docstring to match shannon-oss * docs(worker): adopt richer vuln-renderer docstring from shannon-oss * docs(worker): neutralize billing-detection wording for shannon-oss parity * fix(worker): verify checkpoint hash in the deliverables clone being reset * fix(worker): fail fast on malformed exploitation queue JSON * fix(worker): honor retryable flag when classifying exploitation-queue check failures * fix(worker): fail fast on corrupted session.json in run-scope validation * feat(worker): propagate Temporal cancellation signal into agent and auth pi sessions * fix(worker): mark exploit agent complete when exploitation is skipped so resume skips it * prompts: drop scan description from executive report prompt * refactor(worker): add createGenericSubmitTool for raw JSON-schema submit tools * refactor(worker): gate playwright-cli skill to browser agents via skillsOverride (adopt shannon-oss mechanism) * docs(worker): correct formatLogTime comment to UTC to match toISOString * refactor(worker): converge queue-schemas with shannon-oss (guarded count, decl order) * refactor(worker): converge task-tool with shannon-oss (byte-identical; modelRegistry optional) * fix(worker): use replaceLiteral for all prompt value insertions to prevent $-mangling * fix(worker): classify agent execution failures by error type instead of hardcoding validation * fix(worker): cap auth-failure detail at 250 chars to match shannon-oss * style(worker): apply biome formatting * refactor(worker): remove per-session task delegation cap from task tool * style(cli): collapse usage hint now that the beta tag is gone * chore: mark the pi harness migration as a breaking change BREAKING CHANGE: Google Vertex AI is no longer a supported provider. The CLAUDE_CODE_USE_VERTEX, ANTHROPIC_VERTEX_PROJECT, CLOUD_ML_REGION, and GOOGLE_APPLICATION_CREDENTIALS environment variables, along with the use_vertex, vertex_project, and cloud_ml_region config.toml keys, are removed. Vertex users must switch to Anthropic, AWS Bedrock, or a custom Anthropic-compatible base URL. The CLAUDE_CODE_MAX_OUTPUT_TOKENS environment variable and the max_output_tokens config.toml key are also removed.
232 lines
8.8 KiB
TypeScript
232 lines
8.8 KiB
TypeScript
// Copyright (C) 2025 Keygraph, Inc.
|
|
//
|
|
// This program is free software: you can redistribute it and/or modify
|
|
// it under the terms of the GNU Affero General Public License version 3
|
|
// as published by the Free Software Foundation.
|
|
|
|
/**
|
|
* Deterministic exploit collector → markdown renderer.
|
|
*
|
|
* Single entry point renderExploitDeliverable(vulnClass, state, idToType)
|
|
* covers all 5 exploitation agents (injection, xss, auth, ssrf, authz). The
|
|
* per-class deltas are limited to title and ID prefix; every section, label,
|
|
* and sort rule is class-agnostic. Section headers and bolded field labels
|
|
* mirror the prescribed-Markdown skeleton from the existing exploit-*.txt
|
|
* prompts so downstream report-executive — which reads prose with bolded
|
|
* labels — sees the same structure it sees today.
|
|
*
|
|
* Field-label drift across the 5 prompts ("Evidence of Vulnerability" vs
|
|
* "Why We Believe This Is Vulnerable"; "Attempted Exploitation" vs
|
|
* "What We Tried") is canonicalized here to a single label per field across
|
|
* all classes.
|
|
*
|
|
* Sort order is owned by the renderer:
|
|
* - Successfully Exploited: severity desc (critical → low), then ID asc.
|
|
* - Potential / Validation Blocked: confidence desc (high → low), then ID asc.
|
|
*
|
|
* ## Unprocessed Vulnerabilities surfaces queue IDs the collector did not see —
|
|
* the v1 stand-in for required-call enforcement. The activity passes idToType
|
|
* (queue ID → vulnerability_type, built from queue.json) so each entry renders
|
|
* as `- {ID} ({vulnerability_type})`. Omitted when every queue ID was emitted.
|
|
*/
|
|
|
|
import type { AddExploitInput, VulnClass } from '../collectors/exploit-collector.js';
|
|
|
|
// ============================================================================
|
|
// PER-CLASS CONSTANTS
|
|
// ============================================================================
|
|
|
|
const TITLES: Record<VulnClass, string> = {
|
|
injection: 'Injection Exploitation Evidence',
|
|
xss: 'Cross-Site Scripting (XSS) Exploitation Evidence',
|
|
auth: 'Authentication Exploitation Evidence',
|
|
ssrf: 'SSRF Exploitation Evidence',
|
|
authz: 'Authorization Exploitation Evidence',
|
|
};
|
|
|
|
// ============================================================================
|
|
// SORT ORDER
|
|
// ============================================================================
|
|
|
|
const SEVERITY_ORDER: Record<'critical' | 'high' | 'medium' | 'low', number> = {
|
|
critical: 0,
|
|
high: 1,
|
|
medium: 2,
|
|
low: 3,
|
|
};
|
|
|
|
const CONFIDENCE_ORDER: Record<'high' | 'medium' | 'low', number> = {
|
|
high: 0,
|
|
medium: 1,
|
|
low: 2,
|
|
};
|
|
|
|
type ExploitedEntry = Extract<AddExploitInput, { status: 'exploited' }>;
|
|
type BlockedEntry = Extract<AddExploitInput, { status: 'blocked' }>;
|
|
|
|
function sortExploited(entries: readonly ExploitedEntry[]): ExploitedEntry[] {
|
|
return [...entries].sort((a, b) => {
|
|
const sevDiff = SEVERITY_ORDER[a.severity] - SEVERITY_ORDER[b.severity];
|
|
if (sevDiff !== 0) return sevDiff;
|
|
return a.vulnerability_id.localeCompare(b.vulnerability_id);
|
|
});
|
|
}
|
|
|
|
function sortBlocked(entries: readonly BlockedEntry[]): BlockedEntry[] {
|
|
return [...entries].sort((a, b) => {
|
|
const confDiff = CONFIDENCE_ORDER[a.confidence] - CONFIDENCE_ORDER[b.confidence];
|
|
if (confDiff !== 0) return confDiff;
|
|
return a.vulnerability_id.localeCompare(b.vulnerability_id);
|
|
});
|
|
}
|
|
|
|
// ============================================================================
|
|
// FIELD FORMATTERS
|
|
// ============================================================================
|
|
|
|
function capitalize(value: string): string {
|
|
if (value.length === 0) return value;
|
|
return value[0]!.toUpperCase() + value.slice(1);
|
|
}
|
|
|
|
function renderNumberedList(steps: readonly string[]): string {
|
|
return steps.map((step, idx) => `${idx + 1}. ${step}`).join('\n\n');
|
|
}
|
|
|
|
// ============================================================================
|
|
// PER-FINDING RENDERERS
|
|
// ============================================================================
|
|
|
|
function renderExploitedFinding(entry: ExploitedEntry): string {
|
|
const lines: string[] = [];
|
|
lines.push(`### ${entry.vulnerability_id}: ${entry.title}`);
|
|
lines.push('');
|
|
lines.push('**Summary:**');
|
|
lines.push(`- **Vulnerable location:** ${entry.vulnerable_location}`);
|
|
lines.push(`- **Overview:** ${entry.overview}`);
|
|
lines.push(`- **Impact:** ${entry.impact}`);
|
|
lines.push(`- **Severity:** ${capitalize(entry.severity)}`);
|
|
lines.push('');
|
|
if (entry.prerequisites != null && entry.prerequisites.length > 0) {
|
|
lines.push('**Prerequisites:**');
|
|
lines.push(entry.prerequisites);
|
|
lines.push('');
|
|
}
|
|
lines.push('**Exploitation Steps:**');
|
|
lines.push(renderNumberedList(entry.exploitation_steps));
|
|
lines.push('');
|
|
lines.push('**Proof of Impact:**');
|
|
lines.push(entry.proof_of_impact);
|
|
if (entry.notes != null && entry.notes.length > 0) {
|
|
lines.push('');
|
|
lines.push('**Notes:**');
|
|
lines.push(entry.notes);
|
|
}
|
|
return lines.join('\n');
|
|
}
|
|
|
|
function renderBlockedFinding(entry: BlockedEntry): string {
|
|
const lines: string[] = [];
|
|
lines.push(`### ${entry.vulnerability_id}: ${entry.title}`);
|
|
lines.push('');
|
|
lines.push('**Summary:**');
|
|
lines.push(`- **Vulnerable location:** ${entry.vulnerable_location}`);
|
|
lines.push(`- **Overview:** ${entry.overview}`);
|
|
lines.push(`- **Current Blocker:** ${entry.current_blocker}`);
|
|
lines.push(`- **Potential Impact:** ${entry.potential_impact}`);
|
|
lines.push(`- **Confidence:** ${entry.confidence.toUpperCase()}`);
|
|
lines.push('');
|
|
if (entry.prerequisites != null && entry.prerequisites.length > 0) {
|
|
lines.push('**Prerequisites:**');
|
|
lines.push(entry.prerequisites);
|
|
lines.push('');
|
|
}
|
|
lines.push('**Evidence of Vulnerability:**');
|
|
lines.push(entry.evidence_of_vulnerability);
|
|
lines.push('');
|
|
lines.push('**What We Tried:**');
|
|
lines.push(entry.what_we_tried);
|
|
lines.push('');
|
|
lines.push('**How This Would Be Exploited:**');
|
|
lines.push(renderNumberedList(entry.how_this_would_be_exploited));
|
|
lines.push('');
|
|
lines.push('**Expected Impact:**');
|
|
lines.push(entry.expected_impact);
|
|
if (entry.notes != null && entry.notes.length > 0) {
|
|
lines.push('');
|
|
lines.push('**Notes:**');
|
|
lines.push(entry.notes);
|
|
}
|
|
return lines.join('\n');
|
|
}
|
|
|
|
// ============================================================================
|
|
// SECTION RENDERERS
|
|
// ============================================================================
|
|
|
|
function renderExploitedSection(entries: readonly ExploitedEntry[]): string {
|
|
const heading = '## Successfully Exploited Vulnerabilities';
|
|
if (entries.length === 0) {
|
|
return [heading, '', '*No findings reached a definitive verdict in this category.*'].join('\n');
|
|
}
|
|
const blocks = sortExploited(entries).map(renderExploitedFinding);
|
|
return [heading, '', blocks.join('\n\n')].join('\n');
|
|
}
|
|
|
|
function renderBlockedSection(entries: readonly BlockedEntry[]): string {
|
|
const heading = '## Potential Vulnerabilities (Validation Blocked)';
|
|
if (entries.length === 0) {
|
|
return [heading, '', '*No findings reached a definitive verdict in this category.*'].join('\n');
|
|
}
|
|
const blocks = sortBlocked(entries).map(renderBlockedFinding);
|
|
return [heading, '', blocks.join('\n\n')].join('\n');
|
|
}
|
|
|
|
function renderUnprocessedSection(missingIds: readonly string[], idToType: ReadonlyMap<string, string>): string {
|
|
const heading = '## Unprocessed Vulnerabilities';
|
|
const sortedIds = [...missingIds].sort((a, b) => a.localeCompare(b));
|
|
const lines = sortedIds.map((id) => {
|
|
const type = idToType.get(id);
|
|
return type ? `- ${id} (${type})` : `- ${id}`;
|
|
});
|
|
return [
|
|
heading,
|
|
'',
|
|
'The following queue vulnerabilities did not receive a definitive verdict during this run:',
|
|
'',
|
|
lines.join('\n'),
|
|
].join('\n');
|
|
}
|
|
|
|
// ============================================================================
|
|
// PUBLIC ENTRY POINT
|
|
// ============================================================================
|
|
|
|
export function renderExploitDeliverable(
|
|
vulnClass: VulnClass,
|
|
state: readonly AddExploitInput[],
|
|
idToType: ReadonlyMap<string, string>,
|
|
): string {
|
|
const title = `# ${TITLES[vulnClass]}`;
|
|
|
|
if (state.length === 0 && idToType.size === 0) {
|
|
const body = '*No vulnerabilities were available in the queue for exploitation.*';
|
|
return `${title}\n\n${body}\n`;
|
|
}
|
|
|
|
const exploited = state.filter((e): e is ExploitedEntry => e.status === 'exploited');
|
|
const blocked = state.filter((e): e is BlockedEntry => e.status === 'blocked');
|
|
|
|
const emittedIds = new Set(state.map((e) => e.vulnerability_id));
|
|
const missingIds = [...idToType.keys()].filter((id) => !emittedIds.has(id));
|
|
|
|
const sections: string[] = [title, '', renderExploitedSection(exploited), '', renderBlockedSection(blocked)];
|
|
|
|
if (missingIds.length > 0) {
|
|
sections.push('');
|
|
sections.push(renderUnprocessedSection(missingIds, idToType));
|
|
}
|
|
|
|
return `${sections.join('\n').trimEnd()}\n`;
|
|
}
|