Open a workspace, run Secrets-LE: Detect Secrets, and every detected secret lands in a results document — grouped by file, with line/column positions pointing at the value itself. Run Secrets-LE: Sanitize Secrets to replace the secrets in the active file with a placeholder. Works in VS Code and in VS Code–based editors like Cursor and VSCodium (installable from Open VSX).
Detect hardcoded secrets in source and config. Reports masked previews, never the values.
2026-11-10
ดาว
—
เครื่องมือ
—
อัปไทม์
—
Detection is regex-based over the full text of each file, so it works on any text format — code, configs, .env files, YAML, JSON, logs. It is a pre-commit safety net, not a guarantee: a scanner built on patterns can miss secrets and can flag non-secrets. Review the results.
Reports each finding by type, confidence, key name and 1-based position. Values are never returned — previews are truncated and length-annotated, and the context line has the secret masked out, so a finding can be located without the credential leaving the machine it was found on.
The server takes content and returns data — it reads no files and makes no network requests of its own. Published as secrets-le-mcp on npm and as io.github.nolindnaidoo/secrets-le in the MCP registry.
That prints the tool list and exits — if you see detect_secrets, the server works.
Across a folder or a workspace
Detect reads files from disk and gives one report, grouped by file.
The whole workspace: run Secrets-LE: Detect Secrets from the command palette.
One folder: right-click it in the Explorer and choose Detect Secrets in Folder, or run Secrets-LE: Detect Secrets in Folder and pick one.
# Secrets Detection Results
⚠️ Found 2 potential secret(s):
## 📄 .env (1 secret(s))
### PASSWORD (1)
- Line 1, Column 19
Key: database_password
Type: Password
Confidence: high
Value: hunter2… (14 chars)
Context: DATABASE_PASSWORD=hunter2… (14 chars)
## 📄 src/config.ts (1 secret(s))
### API-KEY (1)
- Line 1, Column 17
Key: apikey
Type: Generic API key
Confidence: high
Value: sk_demo_… (40 chars)
Context: const apiKey = "sk_demo_… (40 chars)";
---
# Warnings
- Not read: dependency folders, build output, caches and lockfiles; images, fonts, archives and other binary files; 1 file(s) ignored by .gitignore. The `secrets-le.workspace.*` settings change this.
What a scan reads. Files come from disk, so an unsaved edit is not seen. A file over the safety size, or one that is not UTF-8 text, is left unread. It stops at 10,000 files or 10,000 listed secrets. The report ends with a line for each thing it left out.
.env files are read even though .gitignore leaves them out. That is where a project keeps its secrets, and it is what this tool is run to look at. scanAlwaysInclude holds **/.env and **/.env.* by default. Empty it to leave them to .gitignore.
A scan that skips is not a clearance. A secret in a dependency folder, in another file .gitignore leaves out, or in a binary file is not reported. To screen everything, turn the three switches off.
What it skips, and how to change that. Three switches are on by default, and each can be turned off on its own in Settings:
Switch
Skips
scanUseDefaultExcludes
Dependency folders, build output, tool caches and lockfiles. The full list is below
scanRespectGitignore
Whatever the project's .gitignore files skip
scanSkipBinaryFiles
Images, fonts, archives and other files that are not text
Two lists adjust the result without turning a switch off. To skip more, add a pattern to scanExcludes. To read something a switch would skip, add it to scanAlwaysInclude:
{
// Also skip the test fixtures.
"secrets-le.workspace.scanExcludes": ["**/fixtures/**"],
// Keep reading .env files, and read the vendored code as well.
"secrets-le.workspace.scanAlwaysInclude": ["**/.env", "**/.env.*", "**/vendor/**"]
}
Secrets-LE: Open Settings opens all of these in the Settings editor.
Not on the list, because they are ordinary folders in many projects: bin, obj, tmp, logs, public, generated. A project that generates those ignores them in git, and the scan reads .gitignore.
In the Problems panel. Turn on secrets-le.workspace.scanProblemsEnabled and each secret is also a warning on its line. Each scan replaces the last one's. A message names the kind of secret, never its value.
The settings that shape a scan are under Settings.
The CLI
The same detection runs from a terminal or a CI step: a Rust CLI in
crate/, sharing one pattern table with the extension
— crate/signatures/patterns.toml —
so the two can never disagree about what counts as a credential. The
same command line also ships on npm as secrets-le, for a machine with
Node and no Rust, and CI holds it to the crate byte for byte on three
operating systems.
secrets-le . # scan a tree
secrets-le --sensitivity high . # only high-confidence findings
secrets-le --no-ignore --hidden . # reach .env and everything git ignores
secrets-le mcp # the same detection over MCP on stdio
Install it from npm (secrets-le) or from crates.io (secrets-le). Both install the same secrets-le command.
npm install -g secrets-le # or run it once: npx secrets-le
cargo install secrets-le
The exit code is the answer: 0 nothing found · 1 findings · 2 the
question was malformed — so secrets-le . is a CI step as it stands.
It never prints a credential. A scanner's output goes into a CI log,
which is archived, often world-readable, and outlives the secret; a
scanner that printed what it found would disclose it more widely than
the commit would have. Previews are capped at eight characters and at
half the value's length, context lines are masked, and there is no flag
that changes either. The extension is the half that can fix what it
finds; the binary only reports.
What gets detected
Thirty-four patterns, in crate/signatures/patterns.toml — the one table
both frontends load.
Category
Types
Named issuers
Anthropic sk-ant-, OpenAI sk-/sk-proj-, xAI xai-, Groq gsk_, Hugging Face hf_, Replicate r8_, GitHub ghp_/gho_/ghu_/ghs_/ghr_/github_pat_, GitLab, Slack xox?- and webhook URLs, Discord webhook URLs, Stripe sk_/rk_ live and test keys and whsec_ webhook secrets, Google AIza… and OAuth client secrets GOCSPX-, Linear lin_api_, DigitalOcean dop_v1_/doo_v1_/dor_v1_, Doppler dp.pt./dp.st., SendGrid, Mailgun, Sentry, npm, PyPI, Docker Hub, HashiCorp Vault, Terraform Cloud, Supabase, Shopify, Square, Azure SAS
Cloud credentials
AWS Access Key IDs (AKIA…, no key name needed), AWS Secret Access Keys, Azure account keys, GCP/Google Cloud keys
Tokens
Generic tokens, bearer tokens, access/refresh tokens, OAuth tokens, JWTs (key-based or bare eyJ… form)
Passwords
password/passwd/pwd values, including compound keys (DATABASE_PASSWORD)
Private keys
Multi-line PEM blocks — RSA/EC, OpenSSH, PGP
Connection data
Database URLs with embedded user:pass@ credentials, connection strings, session IDs, cookies
Key-based patterns accept quoted and unquoted keys, so JSON ("apiKey": "…"), YAML (api_key: …), env (API_KEY=…), and code (apiKey = '…') all match.
Intentional non-detections: template placeholders (${VAR}, {{var}}, <your-key>, xxxxxxxx), version numbers and hostnames that merely look dotted (1.2.3 is not a JWT), GCP project ids (identifiers, not credentials), and database URLs without embedded credentials.
Known limitations: detection is pattern-based — obfuscated, split, or unconventionally named secrets are missed; JWTs whose header isn't standard base64 JSON (eyJ…) are missed; a high-entropy string without a recognizable key name or prefix is not reported.
Commands
Command
Description
Secrets-LE: Detect Secrets
Scan the workspace and open a results document
Secrets-LE: Detect Secrets in Folder
The same for one folder. Also on a folder in the Explorer
Secrets-LE: Sanitize Secrets
Replace detected secrets in the active file (asks for confirmation first)
Secrets-LE: Open Settings
Open Secrets-LE settings
Secrets-LE: Help
Built-in documentation
No command is bound to a key by default. Give any of them one under Keyboard Shortcuts in the editor.
Settings
Setting
Default
Description
secrets-le.detection.sensitivity
medium
low reports everything, medium drops low-confidence matches, high keeps only high-confidence ones
secrets-le.detection.includeApiKeys
true
Detect API keys and cloud credentials
secrets-le.detection.includePasswords
true
Detect passwords
secrets-le.detection.includeTokens
true
Detect tokens and JWTs
secrets-le.detection.includePrivateKeys
true
Detect PEM private-key blocks
secrets-le.sanitization.replaceWith
***REDACTED***
Replacement text used by Sanitize
secrets-le.workspace.scanPatterns
["**/*"]
The files a folder or workspace scan reads
secrets-le.workspace.scanUseDefaultExcludes
true
Skip dependency folders, build output, caches and lockfiles
secrets-le.workspace.scanRespectGitignore
true
Skip what the project's .gitignore files skip
secrets-le.workspace.scanSkipBinaryFiles
true
Skip images, fonts, archives and other files that are not text
secrets-le.workspace.scanExcludes
[]
More files to skip, as glob patterns
secrets-le.workspace.scanAlwaysInclude
["**/.env", "**/.env.*"]
Files to read even when one of the three above would skip them
secrets-le.workspace.scanMaxFiles
10000
The most files one scan reads
secrets-le.workspace.scanMaxResults
10000
The most secrets one scan lists before it stops reading
secrets-le.workspace.scanProblemsEnabled
false
Also show the secrets a scan finds in the Problems panel
secrets-le.safety.enabled
true
Guardrails for very large files
secrets-le.safety.fileSizeWarnBytes
1000000
Skip/refuse files above this size
secrets-le.dedupeEnabled
false
Collapse identical value+type detections in results
secrets-le.showPositions
true
Show the line and column of each secret found
secrets-le.copyToClipboardEnabled
false
Also copy results to the clipboard
secrets-le.clipboardIncludesPositions
true
Include the line and column in that copy
secrets-le.openResultsSideBySide
true
Open results beside the current editor
secrets-le.notificationsLevel
important
all = every notification, important = warnings + errors, silent = errors only
secrets-le.statusBar.enabled
true
Show the status bar item
secrets-le.telemetryEnabled
false
Local-only event log (see Privacy)
Languages
Twelve languages besides English:
German · Spanish · French · Indonesian · Italian · Japanese · Korean ·
Portuguese (Brazil) · Russian · Ukrainian · Vietnamese · Chinese (Simplified)
Both halves are covered — the manifest (command titles, setting names and
descriptions) and everything shown while the extension runs (notifications,
the status bar, quick-picks and prompts). The extension follows VS Code's
display language, so it matches whatever the editor is already set to; no
setting of its own.
Privacy & security
No network access. The extension never sends data anywhere. The telemetryEnabled setting only writes events to a local Output Channel you can inspect (Secrets-LE Telemetry).
The MCP server never returns a secret. Its output goes to whatever model called it, so previews are truncated and length-annotated and the surrounding context line is masked, using the same utils/mask helpers as the report. There is no argument that turns this off, and the bundle gate fails the build if a value ever appears in a response — verified by making the tool leak on purpose and watching the gate catch it.
Error notifications redact home directories and credential-shaped fragments before display.
Sanitize always asks for confirmation before editing your file, and edits are normal undo-able document edits.
One rating prompt, at most twice. On the 3rd successful use the extension asks once whether you would rate it, and once more on the 20th if you chose Later or dismissed it. Don't Ask Again ends it. Setting notificationsLevel to important or silent yourself turns it off. The counts are kept in VS Code's extension storage and nothing is sent anywhere; Rate opens the listing you installed from — the VS Code Marketplace or Open VSX — in your browser.
Documentation
What
Where
What the tool is allowed to say — scope, output contract, refusals, non-goals
Median of 7 runs after warmup, on Apple M5 Pro, 24 GB RAM, Node 24.3.0. Inputs are generated
by scripts/benchmark.ts rather than checked in, so the sizes above are
exactly what was measured. Reproduce with bun run benchmark.
These are machine-specific and are not asserted in CI — a benchmark that gates
a build only tells you how busy the runner was.
Testing
Metric
Coverage
Statements
86.89%
Branches
78.60%
Functions
86.45%
Lines
88.49%
412 test cases across 24 files, plus an integration suite that runs
in a real VS Code extension host and an end-to-end test that installs the
built .vsix into a clean profile.
Generated from a real run — coverage/coverage-summary.json and
coverage/test-results.json — by scripts/coverage-readme.js; CI fails if
this section drifts. Reproduce with bun run test:coverage, and the case
count is the one vitest prints.
More from the LE family
Sixteen single-purpose tools for the work in front of every model. Each ships
a Rust CLI and an MCP server. One page: letools.dev
Get it out
String-LE — Extract every string in a codebase, so a person can read them
Numbers-LE — Extract every hardcoded number in a codebase, so a person can check them
Units-LE — Extract every quantity with its unit, normalized, and refuse the ambiguous ones by name
Dates-LE — Extract every date and timestamp, and the exact instant each one resolves to
IDs-LE — Extract every UUID, ULID, NanoID, ObjectId and Snowflake, and decode the time inside
IPs-LE — Extract every IP address, CIDR block and MAC, normalized and classified by scope
URLs-LE — Extract every URL in a codebase, with its protocol and exact position
Paths-LE — Extract every file path in a codebase, and say whether it still points at anything
Colors-LE — Extract every color in a codebase, and say which ones are not in your palette
Check it
Regex-LE — Find every regex in a codebase, and report which can be driven into catastrophic backtracking
Versions-LE — Find where one dependency is constrained differently across a repository's manifests
i18n-LE — Identify the i18n library a project uses, then audit its catalogs by that library's rules
Scrape-LE — Check whether a page is scrapeable before the scraper is written, and say when it cannot tell
Guard it
Secrets-LE — Find hardcoded credentials in a codebase, and never print one into the report
EnvSync-LE — Compare the dotenv files in a tree, and say which keys are missing from which
Unicode-LE — Find the Unicode that hides meaning — bidi controls, invisibles, homoglyphs, mixed scripts
Each stands on its own: no shared crate, no published core. Where two of them
agree, it is because the same answer was right twice.
Rust — pixelcoords and pixelactions are one loop: pixelcoords answers
where, pixelactions acts there. Their own tools, their own voice — not
part of the LE family.