feat(DIS-102): sanitizeForDiscord strips paths and tokens before Discord send #46

Merged
dev merged 4 commits from phase-1/sanitize-for-discord into main 2026-04-10 08:05:44 +00:00
5 changed files with 167 additions and 4 deletions

View file

@ -3,6 +3,7 @@ import * as path from "path";
import * as YAML from "yaml"; import * as YAML from "yaml";
import { ZodError } from "zod"; import { ZodError } from "zod";
import { AgentYamlSchema } from "./schema"; import { AgentYamlSchema } from "./schema";
import type { AgentYaml as AgentIdentity } from "./schema";
export type { AgentYaml as AgentIdentity } from "./schema"; export type { AgentYaml as AgentIdentity } from "./schema";

View file

@ -6,6 +6,7 @@ import * as YAML from "yaml";
import { ZodError } from "zod"; import { ZodError } from "zod";
import { childLogger } from "../runtime/logger"; import { childLogger } from "../runtime/logger";
import { DisclawConfigSchema } from "./schema"; import { DisclawConfigSchema } from "./schema";
import type { DisclawConfig } from "./schema";
export type { DisclawConfig } from "./schema"; export type { DisclawConfig } from "./schema";

View file

@ -1,7 +1,9 @@
import * as os from "os";
import { Message, TextChannel } from "discord.js"; import { Message, TextChannel } from "discord.js";
import { DisclawDatabase } from "./db/database"; import { DisclawDatabase } from "./db/database";
import { DisclawConfig } from "./config/loader"; import { DisclawConfig } from "./config/loader";
import { runAgent } from "./agent/runner"; import { runAgent } from "./agent/runner";
import { sanitizeForDiscord } from "./runtime/sanitize";
const DISCORD_MAX_LENGTH = 2000; const DISCORD_MAX_LENGTH = 2000;
@ -105,8 +107,18 @@ export async function routeMessage(
message.author.username message.author.username
); );
// Sanitize the full response once, then use the sanitized version
// for both Discord output and conversation history storage.
// This prevents a feedback loop where raw paths in stored history
// get fed back to the agent as context, causing it to reproduce them.
const sanitizeOpts = {
repoRoot: workspace.workspace_path,
home: os.homedir(),
};
const sanitizedResponse = sanitizeForDiscord(response, sanitizeOpts);
// Split and send response // Split and send response
const chunks = splitMessage(response); const chunks = splitMessage(sanitizedResponse);
let firstMsgId: string | null = null; let firstMsgId: string | null = null;
for (const chunk of chunks) { for (const chunk of chunks) {
@ -116,19 +128,23 @@ export async function routeMessage(
} }
} }
// Store assistant response (use first message id for dedup) // Store sanitized assistant response (use first message id for dedup)
if (firstMsgId) { if (firstMsgId) {
db.addConversation( db.addConversation(
workspace.id, workspace.id,
firstMsgId, firstMsgId,
"assistant", "assistant",
response, sanitizedResponse,
null null
); );
} }
} catch (error) { } catch (error) {
const msg = error instanceof Error ? error.message : "Unbekannter Fehler"; const msg = error instanceof Error ? error.message : "Unbekannter Fehler";
await channel.send(`Fehler bei der Verarbeitung: ${msg}`).catch(() => {}); const safeMsg = sanitizeForDiscord(`Fehler bei der Verarbeitung: ${msg}`, {
repoRoot: workspace.workspace_path,
home: os.homedir(),
});
await channel.send(safeMsg).catch(() => {});
} finally { } finally {
clearInterval(typingInterval); clearInterval(typingInterval);
} }

58
src/runtime/sanitize.ts Normal file
View file

@ -0,0 +1,58 @@
export interface SanitizeOptions {
repoRoot?: string; // e.g. /home/user/.disclaw
home?: string; // os.homedir()
}
/**
* Sanitizes agent output before it is sent to a Discord channel.
* Replaces absolute paths and sensitive tokens so they are never leaked.
*
* Replacement order is intentional longest/most-specific patterns first:
* 1. repoRoot <disclaw>
* 2. home ~
* 3. Discord bot token pattern <redacted>
* 4. Anthropic API key pattern <redacted>
*/
/** Returns both backslash and forward-slash variants of a path (for Windows). */
function pathVariants(p: string): string[] {
const fwd = p.replace(/\\/g, "/");
const bwd = p.replace(/\//g, "\\");
return fwd === bwd ? [p] : [p, fwd, bwd];
}
function replaceAll(text: string, search: string, replacement: string): string {
return text.split(search).join(replacement);
}
export function sanitizeForDiscord(
text: string,
opts: SanitizeOptions = {}
): string {
let result = text;
// 1. Replace repo root before home so that sub-paths are caught by the
// more specific substitution first (repoRoot is usually inside home).
if (opts.repoRoot) {
for (const variant of pathVariants(opts.repoRoot)) {
result = replaceAll(result, variant, "<disclaw>");
}
}
// 2. Replace home directory with ~ (both slash styles for Windows)
if (opts.home) {
for (const variant of pathVariants(opts.home)) {
result = replaceAll(result, variant, "~");
}
}
// 3. Discord bot token — format: [MN]<23 base64url>.<6 base64url>.<27 base64url>
result = result.replace(
/[MN][A-Za-z\d]{23}\.[\w-]{6}\.[\w-]{27}/g,
"<redacted>"
);
// 4. Anthropic API key — format: sk-ant-<20+ base64url chars>
result = result.replace(/sk-ant-[A-Za-z0-9\-_]{20,}/g, "<redacted>");
return result;
}

View file

@ -0,0 +1,87 @@
import { describe, it, expect } from "vitest";
import { sanitizeForDiscord } from "../../src/runtime/sanitize";
describe("sanitizeForDiscord", () => {
// (a) Repo-root path is replaced with <disclaw>
it("replaces repoRoot with <disclaw>", () => {
const text = "Reading config from /home/user/project/disclaw.yaml";
const result = sanitizeForDiscord(text, { repoRoot: "/home/user/project" });
expect(result).toBe("Reading config from <disclaw>/disclaw.yaml");
expect(result).not.toContain("/home/user/project");
});
// (b) Home directory is replaced with ~
it("replaces home directory with ~", () => {
const text = "Workspace located at /home/user/.disclaw/workspaces/agent-x";
const result = sanitizeForDiscord(text, { home: "/home/user" });
expect(result).toBe("Workspace located at ~/.disclaw/workspaces/agent-x");
expect(result).not.toContain("/home/user");
});
// (c) Discord bot token pattern is redacted
it("redacts Discord bot tokens", () => {
const token = "MTA1NjYwNDczNTA3NDU2NjE2.GbXkRa.abcdefghijklmnopqrstuvwxyz123";
const text = `Bot initialized with token ${token} — ready.`;
const result = sanitizeForDiscord(text);
expect(result).not.toContain(token);
expect(result).toContain("<redacted>");
});
// (d) Anthropic API key is redacted
it("redacts Anthropic API keys", () => {
const key = "sk-ant-api03-ABCDEFGHIJ1234567890abcdefghij";
const text = `Using key ${key} for requests.`;
const result = sanitizeForDiscord(text);
expect(result).not.toContain(key);
expect(result).toContain("<redacted>");
});
// (e) Harmless text is not modified (no false positives)
it("does not modify harmless text", () => {
const text = "Hello! The agent processed your request successfully in 120ms.";
const result = sanitizeForDiscord(text, { home: "/home/user", repoRoot: "/home/user/project" });
expect(result).toBe(text);
});
// (f) Empty string returns empty string
it("handles empty string input", () => {
expect(sanitizeForDiscord("")).toBe("");
expect(sanitizeForDiscord("", { home: "/home/user", repoRoot: "/home/user/project" })).toBe("");
});
// (g) No opts provided — plain text stays unchanged when no token patterns present
it("returns text unchanged when no opts and no token patterns", () => {
const text = "This is a perfectly normal response from the agent.";
expect(sanitizeForDiscord(text)).toBe(text);
});
// Ordering: repoRoot (sub-path of home) is replaced before home
it("replaces repoRoot before home so nested paths are handled correctly", () => {
const text = "/home/user/project/src/index.ts and /home/user/other.ts";
const result = sanitizeForDiscord(text, {
repoRoot: "/home/user/project",
home: "/home/user",
});
expect(result).toBe("<disclaw>/src/index.ts and ~/other.ts");
});
// Windows-style backslash paths are sanitized
it("replaces Windows backslash paths", () => {
const text = "File at C:\\Users\\dev\\.disclaw\\workspaces\\agent-x\\src\\main.ts";
const result = sanitizeForDiscord(text, {
repoRoot: "C:\\Users\\dev\\.disclaw\\workspaces\\agent-x",
home: "C:\\Users\\dev",
});
expect(result).toBe("File at <disclaw>\\src\\main.ts");
});
// Mixed slash styles (common on Windows when tools output forward slashes)
it("replaces forward-slash variants of Windows paths", () => {
const text = "File at C:/Users/dev/.disclaw/workspaces/agent-x/src/main.ts";
const result = sanitizeForDiscord(text, {
repoRoot: "C:\\Users\\dev\\.disclaw\\workspaces\\agent-x",
home: "C:\\Users\\dev",
});
expect(result).toBe("File at <disclaw>/src/main.ts");
});
});