check-silent-bypass
Review recently-changed TypeScript in `packages/` for SILENT BYPASS patterns — code that looks normal but skips a load-bearing security or correctness check…
Skill by motxx
Chalk Labs
Need help deploying check-silent-bypass?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/motxx-check-silent-bypass
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: backend, security, writing, finance
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/check-silent-bypass
What's inside
- Skill name
check-silent-bypass- Description
- Review recently-changed TypeScript in `packages/` for SILENT BYPASS patterns — code that looks normal but skips a load-bearing security or correctness check on a plausible-looking branch. **Invoke proactively** after editing files that touch verification, validation, settlement, redemption, authentication, authorization, escrow, payment, signing, or quorum logic, and before committing changes to those areas. Also invoke when the user says "silent bypass check", "silent bypass", "/check-silent-bypass", "サイレントバイパスチェック", or "バイパスチェック". Deterministic lints catch explicit stub markers; this skill catches semantic bypasses by reading code. Runs inside the existing session; no API key required.
- License
- MIT
- Instructions
- 205 lines in SKILL.md · view on GitHub
Agent Skills spec issues
- files: referenced but missing: scripts/silent-bypass-verify.ts
Files (1)
- SKILL.md
SKILL.md (first lines)
# Silent-Bypass Review
Read the in-scope `.ts` / `.tsx` files, apply the rubric below, and report
findings inline. Do NOT modify any files — this is read-only review.
## Scope
- **Include:** `packages/<pkg>/src/**/*.ts(x)`
- **Exclude:** `*.test.ts(x)`, `*.d.ts`, `packages/sdk/src/testing/`,
`dist/`, `node_modules/`
UI, examples, and scripts are out of scope: silent bypass is a trust
enforcement concern, and trust is enforced in package code. Client-side
validation is a UX hint, not a security control.
File selection, in order:
1. If the user supplies paths or globs as arguments, scan exactly those
(still respecting the exclude list).
2. If files in scope were edited in the current session, scan those.
3. Otherwise scan in-scope files from `git diff --name-only
origin/main...HEAD`.
4. Otherwise scan unstaged in-scope files from `git diff --name-only HEAD`.
If nothing is in scope, say so and stop — do NOT scan the whole repository
unless the user explicitly asks.
## How to read each file
1. Read the **full** file — naming and return-type analysis need
surrounding context, not just the diff hunk.
2. Walk the body looking for the three patterns below.
3. For each candidate, decide: real bypass, or a documented negative case?
4. Lines ending with `// allow-bypass: <reason>` are author-acknowledged —
drop those candidates.
## Rubric — three patterns
### Pattern A — Trust boundary skipped on a plausible branch
A real validation path is gated behind a condition that *looks* like a
sanity check (well-formed input, expected prefix, common case), and the
other branch silently accepts the input without performing the check.
**Flag:**
```ts
if (token.startsWith("cashuB")) {
const decoded = verify(token); // only well-formed tokens get verified
return { valid: decoded.ok };
}
return { valid: true }; // anything else passes
```
```ts
if (env.NODE_ENV === "production") {
await checkRateLimit(req);
}
return handler(req); // dev/test/staging skip the limit
```
**Do NOT flag:**
- Validation whose other branch returns a clear error
(`return { ok: false, error: "..." }`).
- Type narrowing whose other branch is a compile-time impossibility
(exhaustive switch with `assertNever`).
- Dispatch tables where each branch fully handles its own type.
### Pattern B — Error swallowed and converted to success
A try/catch wraps a load-bearing operation; the catch discards the error
and returns a success-shaped value, or the function reports success when
the operation provably failed.
**Flag:**
```ts
try {
const decoded = getDecodedToken(token);
return { ok: true, decoded };
} catch {
return { ok: true }; // swallows the decode failure
}
```
```ts
const result = await db.update(...);
return { updated: true }; // ignores result.rowsAffected
```
**Do NOT flag:**
- A catch that logs and returns a clearly-failed result.
- A catch that re-throws or wraps the error.
- A catch handling one specific narrow error type and re-throwing the rest.
- Best-effort operations whose failure is documented as acceptable.
### Pattern C — Success claimed without doing the work
A function whose name implies a side effect (`settle`, `verify`,
`validate`, `persist`, `send`, `redeem`, `finalize`, `commit`, …) returns a
success-shaped value without performing it.
**Flag:**
```ts
async settle(token: string): Promise<{ settled: boolean }> {
return { settled: true }; // name claims work, body does none
}
```
**Do NOT flag:**
- Functions named `is*`, `has*`, `can*`, `should*`, `get*`, `from*`,
`format*`, `build*`, `compute*` — read-only by name.
- Stubs returning a clearly-failed result with an explicit error — loud
failure is the correct surface.
- Pass-through wrappers that delegate.From motxx/anchr, MIT.
More from motxx/anchr
- 0.524score
resolve-issues
motxx
Resolve repository-tracked issues from `docs/issues/pending/`.
SkillBFreshness: BClaude Code +9 - 0.521score
arch-lint-llm
motxx
Review TypeScript architecture for semantic ownership violations that static lint cannot detect.
SkillBFreshness: BClaude Code +9 - 0.516score
plain-terms
motxx
Review technical prose or identifiers for coined compound terms that hide the concrete subject, artifact, or behavior, especially wording produced or…
SkillBFreshness: BClaude Code +9 - 0.474score
make-issues
motxx
Create repository-tracked Markdown issues under `docs/issues/pending/` — new top-level issues, or child issues that split an existing parent.
SkillCFreshness: CClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Jun 12, 2026 (4 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/motxx/anchr
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from motxx
- 0.524score
resolve-issues
motxx
Resolve repository-tracked issues from `docs/issues/pending/`.
SkillBFreshness: BClaude Code +9 - 0.521score
arch-lint-llm
motxx
Review TypeScript architecture for semantic ownership violations that static lint cannot detect.
SkillBFreshness: BClaude Code +9 - 0.516score
plain-terms
motxx
Review technical prose or identifiers for coined compound terms that hide the concrete subject, artifact, or behavior, especially wording produced or…
SkillBFreshness: BClaude Code +9 - 0.474score
make-issues
motxx
Create repository-tracked Markdown issues under `docs/issues/pending/` — new top-level issues, or child issues that split an existing parent.
SkillCFreshness: CClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get check-silent-bypass running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
check-silent-bypass skill. Agent Mart. https://agentmart.dev/skills/motxx-check-silent-bypass (last updated 2026-10-24)