fix(security): redact secret patterns from CLI log and error output#672
Open
WuKongAI-CMU wants to merge 1 commit intoNVIDIA:mainfrom
Open
fix(security): redact secret patterns from CLI log and error output#672WuKongAI-CMU wants to merge 1 commit intoNVIDIA:mainfrom
WuKongAI-CMU wants to merge 1 commit intoNVIDIA:mainfrom
Conversation
Add a redact() helper to bin/lib/runner.js that masks known secret patterns before they reach CLI error messages. Covers: - NVIDIA API keys (nvapi-*) - NVCF keys (nvcf-*) - Bearer tokens - Generic key/token/password assignments Applied to all run() and runInteractive() error output where the failing command string is logged. Non-secret strings pass through unchanged. Fixes NVIDIA#664 Signed-off-by: peteryuqin <peter.yuqin@gmail.com>
📝 WalkthroughWalkthroughThis PR adds a secret redaction utility to prevent API keys and tokens from leaking in CLI error output. A Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~12 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
bin/lib/runner.js (1)
46-58:⚠️ Potential issue | 🟡 MinorApply
redact()torunCaptureerror paths for consistency with other run functions.
runCapturethrows unredacted exceptions (line 57), whilerun()andrunInteractive()both applyredact()to command failures. This inconsistency creates a defensive programming gap: although current callers catch errors without logging, future modifications could expose the unredacted command string containing secrets. Align with the existing pattern by wrapping the thrown error message withredact().🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@bin/lib/runner.js` around lines 46 - 58, The catch block in runCapture currently rethrows the raw error; update runCapture so that in the catch, after handling opts.ignoreError, it rethrows the error wrapped with redact() (i.e., throw redact(err)) to match the behavior of run() and runInteractive(); keep the existing ignoreError return "" behavior and ensure you reference the runCapture function and redact() wrapper when making the change.
🧹 Nitpick comments (2)
test/runner.test.js (1)
144-182: Consider adding test cases for edge patterns and potential bypasses.The test suite covers the happy paths well. Consider adding tests for:
- Case sensitivity: Does
BEARER(uppercase) orbearer(lowercase) get masked? (The regex usesgiflag, so it should)- Quoted values:
API_KEY="secret123abc"— does the quote handling work?- Multiple secrets in one string:
nvapi-key1 nvapi-key2Optional: Additional test cases
it("masks bearer tokens case-insensitively", () => { const { redact } = require(runnerPath); expect(redact("authorization: bearer someToken1234")).toContain("****"); }); it("masks multiple secrets in one string", () => { const { redact } = require(runnerPath); const input = "nvapi-firstkey12345 and nvapi-secondkey6789"; const output = redact(input); expect(output).not.toContain("firstkey"); expect(output).not.toContain("secondkey"); });🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@test/runner.test.js` around lines 144 - 182, Add unit tests to cover edge patterns and potential bypasses for the redact function: include case-insensitive bearer variations (e.g., "BEARER"/"bearer") to ensure redact (from runnerPath) handles regex flags correctly, add quoted assignment cases (e.g., API_KEY="secret123abc") to verify quotes are masked, and add a test with multiple secrets in one string (e.g., "nvapi-first nvapi-second") to confirm redact masks all occurrences; place these new it() cases alongside the existing describe("redact") block referencing redact from runnerPath.bin/lib/runner.js (1)
66-71: Add coverage for_KEY=andghp_patterns to match shell script.The JS patterns miss two variants that are covered in
scripts/debug.sh(lines 107-113):
ghp_prefix (GitHub personal access tokens)_KEY=suffix (catchesNVIDIA_API_KEY,OTHER_API_KEY, etc.)The shell script redacts both, but the JS version only covers
API_KEY,SECRET,TOKEN,PASSWORD,CREDENTIAL.Suggested update
const SECRET_PATTERNS = [ /nvapi-[A-Za-z0-9_-]{10,}/g, /nvcf-[A-Za-z0-9_-]{10,}/g, + /ghp_[A-Za-z0-9]{30,}/g, /(?<=Bearer\s)[A-Za-z0-9_.+/=-]{10,}/gi, - /(?<=(?:API_KEY|SECRET|TOKEN|PASSWORD|CREDENTIAL)[=: ]['"]?)[A-Za-z0-9_.+/=-]{10,}/gi, + /(?<=(?:_KEY|API_KEY|SECRET|TOKEN|PASSWORD|CREDENTIAL)[=: ]['"]?)[A-Za-z0-9_.+/=-]{10,}/gi, ];🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@bin/lib/runner.js` around lines 66 - 71, SECRET_PATTERNS in runner.js is missing coverage for GitHub PATs and env vars that end with _KEY; update the SECRET_PATTERNS array (the constant named SECRET_PATTERNS) to add a regex for the ghp_ prefix (e.g. match ghp_[A-Za-z0-9_-]{10,}) and extend the existing lookbehind group that matches API_KEY/SECRET/TOKEN/PASSWORD/CREDENTIAL to also accept variable names that end with _KEY (so include _KEY in that lookbehind), keeping the same overall length threshold and flags as the other patterns.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Outside diff comments:
In `@bin/lib/runner.js`:
- Around line 46-58: The catch block in runCapture currently rethrows the raw
error; update runCapture so that in the catch, after handling opts.ignoreError,
it rethrows the error wrapped with redact() (i.e., throw redact(err)) to match
the behavior of run() and runInteractive(); keep the existing ignoreError return
"" behavior and ensure you reference the runCapture function and redact()
wrapper when making the change.
---
Nitpick comments:
In `@bin/lib/runner.js`:
- Around line 66-71: SECRET_PATTERNS in runner.js is missing coverage for GitHub
PATs and env vars that end with _KEY; update the SECRET_PATTERNS array (the
constant named SECRET_PATTERNS) to add a regex for the ghp_ prefix (e.g. match
ghp_[A-Za-z0-9_-]{10,}) and extend the existing lookbehind group that matches
API_KEY/SECRET/TOKEN/PASSWORD/CREDENTIAL to also accept variable names that end
with _KEY (so include _KEY in that lookbehind), keeping the same overall length
threshold and flags as the other patterns.
In `@test/runner.test.js`:
- Around line 144-182: Add unit tests to cover edge patterns and potential
bypasses for the redact function: include case-insensitive bearer variations
(e.g., "BEARER"/"bearer") to ensure redact (from runnerPath) handles regex flags
correctly, add quoted assignment cases (e.g., API_KEY="secret123abc") to verify
quotes are masked, and add a test with multiple secrets in one string (e.g.,
"nvapi-first nvapi-second") to confirm redact masks all occurrences; place these
new it() cases alongside the existing describe("redact") block referencing
redact from runnerPath.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 29d3d3d1-7a52-4a7e-a4ec-c0af1bf967dd
📒 Files selected for processing (2)
bin/lib/runner.jstest/runner.test.js
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a
redact()helper tobin/lib/runner.jsthat masks known secret patterns before they reach CLI error messages.Patterns covered:
nvapi-*)nvcf-*)Applied to all
run()andrunInteractive()error output where the failing command string is logged. Non-secret strings pass through unchanged.Test plan
test/runner.test.jscovering all pattern types + non-secret passthrough + non-string inputFixes #664
Summary by CodeRabbit
Bug Fixes
Tests