CtrlK
BlogDocsLog inGet started
Tessl Logo

base/typescript-engineering

Use this skill whenever the user asks you to write, edit, review, refactor, debug, or design TypeScript or TSX code. It is especially relevant for application code, backend routes, React/UI work, schemas, runtime boundaries, persistence, async workflows, API contracts, tests, lint/typecheck fixes, and code review. Apply it even when the user does not explicitly mention "TypeScript" if the files or project are TypeScript-based.

89

1.26x
Quality

85%

Does it follow best practices?

Impact

95%

1.26x

Average score across 5 eval scenarios

SecuritybySnyk

Passed

No known issues

Overview
Quality
Evals
Security
Files

criteria.jsonevals/scenario-1/

{
  "context": "Tests whether the agent hardens a TypeScript HTTP boundary while preserving ownership, contracts, security, and focused proof. The starter route has unchecked inputs, unsafe SQL, nullable data leakage, weak naming, and missing tests.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Existing conventions",
      "description": "Reads and follows the provided local project instructions and nearby conventions, including named exports and the existing validation dependency.",
      "max_score": 8
    },
    {
      "name": "Boundary schema",
      "description": "Adds a runtime schema for the incoming request payload at the route or domain boundary, and infers the TypeScript type from that schema rather than duplicating it by hand.",
      "max_score": 10
    },
    {
      "name": "Parse once",
      "description": "Parses untrusted data once at the boundary and passes a trusted typed value inward without repeated validation or defensive optional chaining in domain code.",
      "max_score": 8
    },
    {
      "name": "Unknown narrowed",
      "description": "Uses unknown or schema parsing for boundary data and does not use any or broad type assertions for the payload.",
      "max_score": 8
    },
    {
      "name": "Null normalized",
      "description": "Normalizes external null values to internal undefined for optional notes or coupon fields while preserving any externally meaningful raw value where audit data is stored.",
      "max_score": 8
    },
    {
      "name": "Thin route",
      "description": "Keeps framework glue thin by moving meaningful order-fulfillment logic into named domain-owned functions near the order feature rather than a generic utils or services file.",
      "max_score": 8
    },
    {
      "name": "Typed outcomes",
      "description": "Uses a local discriminated union or equivalent named outcome for expected domain results such as duplicate event, invalid order, or fulfilled order instead of returning null or undefined as failure signals.",
      "max_score": 10
    },
    {
      "name": "Secure query",
      "description": "Uses the provided query builder or parameterized database call and does not concatenate request data into SQL.",
      "max_score": 8
    },
    {
      "name": "Auth nearby",
      "description": "Performs webhook signature or tenant authorization checks close to the protected fulfillment operation and fails closed when the signature or tenant is missing or ambiguous.",
      "max_score": 8
    },
    {
      "name": "Domain constants",
      "description": "Moves business-rule literals such as retry count, timeout, event name, or maximum line items into an owner-level constants file with meaning-based names.",
      "max_score": 6
    },
    {
      "name": "Focused tests",
      "description": "Adds focused tests for parsing, null normalization, unsafe inputs, duplicate-event behavior, and query safety without relying on broad snapshots.",
      "max_score": 10
    },
    {
      "name": "Check evidence",
      "description": "Includes a short change note that states what deterministic checks were run, or if a check could not run, what was manually verified and what risk remains.",
      "max_score": 8
    }
  ]
}

evals

scenario-1

criteria.json

task.md

SKILL.md

tile.json