CtrlK
BlogDocsLog inGet started
Tessl Logo

tessl-labs/skill-optimizer

Optimize your skills and tiles: review SKILL.md quality, generate eval scenarios, run evals, compare across models, diagnose gaps, and re-run until scores improve.

88

1.07x
Quality

94%

Does it follow best practices?

Impact

88%

1.07x

Average score across 24 eval scenarios

SecuritybySnyk

Passed

No known issues

Overview
Quality
Evals
Security
Files

task.mdevals/scenario-15/

Bundle File Audit

Problem Description

A developer has a skill tile that's grown over time. They've noticed the tile size is larger than expected and suspect there might be files in the bundle that are no longer referenced. They want to audit their skill bundle to identify any orphaned files that could be removed to reduce bloat, or should be linked if they contain valuable content.

Your job is to analyze the skill bundle below, identify which files are referenced vs orphaned, and provide recommendations for each orphaned file.

Output Specification

Produce a file bundle_audit.md containing:

  • A "Bundle Contents" section listing all files in the bundle
  • A "Referenced Files" section listing files that are linked from SKILL.md
  • An "Orphaned Files" section listing files that exist but are never referenced
  • A "Recommendations" section with specific suggestions for each orphaned file (link with routing signals OR remove with justification)

Input Files

The following files represent the skill bundle. Extract them before beginning.

=============== FILE: database_helper/SKILL.md ===============

name: database-helper description: | Helps with database operations including migrations, queries, and connection pooling. Use when working with databases, running migrations, writing SQL queries, or managing database connections.

Database Helper

Simplify database operations with migration management, query building, and connection pooling.

Quick Start

from db_helper import Database

db = Database(url="postgresql://localhost/mydb")
result = db.query("SELECT * FROM users WHERE active = true")

Migrations

See MIGRATIONS.md for creating, running, and rolling back database migrations.

Connection Pooling

Connection pooling is enabled by default with sensible defaults. See POOLING.md for pool sizing, timeout configuration, and connection lifecycle management.

Query Building

Use the query builder for complex queries:

query = db.select("users").where("active", True).order_by("created_at")
results = query.execute()

Notes

  • Always use parameterized queries to prevent SQL injection
  • Connection pool defaults to 10 connections =============== END FILE ===============

=============== FILE: database_helper/MIGRATIONS.md ===============

Migrations Guide

How to create and run database migrations.

Creating Migrations

db-helper migrate create add_users_table

Running Migrations

db-helper migrate up

=============== END FILE ===============

=============== FILE: database_helper/POOLING.md ===============

Connection Pooling

Configuration options for the connection pool.

  • pool_size: Maximum number of connections (default: 10)
  • pool_timeout: Timeout in seconds (default: 30)
  • pool_recycle: Recycle connections after N seconds (default: 3600) =============== END FILE ===============

=============== FILE: database_helper/TRANSACTIONS.md ===============

Transaction Management

How to use transactions for atomic operations.

Basic Transactions

with db.transaction():
    db.execute("UPDATE accounts SET balance = balance - 100 WHERE id = 1")
    db.execute("UPDATE accounts SET balance = balance + 100 WHERE id = 2")

Nested Transactions

Use savepoints for nested transactions:

with db.transaction():
    db.execute("INSERT INTO orders ...")
    with db.savepoint():
        db.execute("INSERT INTO order_items ...")

=============== END FILE ===============

=============== FILE: database_helper/PERFORMANCE.md ===============

Performance Tips

Optimization strategies for database operations.

Indexes

Create indexes on frequently queried columns:

CREATE INDEX idx_users_email ON users(email);

Batch Operations

Use batch inserts for bulk data:

db.batch_insert("users", [
    {"name": "Alice", "email": "alice@example.com"},
    {"name": "Bob", "email": "bob@example.com"}
])

=============== END FILE ===============

=============== FILE: database_helper/SECURITY.md ===============

Security Best Practices

Database security guidelines.

  • Always use parameterized queries
  • Never store passwords in plaintext
  • Use SSL/TLS for connections in production
  • Rotate credentials regularly
  • Apply principle of least privilege =============== END FILE ===============

=============== FILE: database_helper/LEGACY_EXAMPLES.md ===============

Legacy Examples

Old examples from version 1.x (deprecated).

These examples use the old API and should not be used. Kept for reference during migration.

# OLD API - DO NOT USE
db = DBConnection(host="localhost", port=5432)
db.connect()
db.raw_query("SELECT * FROM users")
db.disconnect()

=============== END FILE ===============

=============== FILE: database_helper/DRAFT_REPLICATION.md ===============

Replication Setup (Draft)

THIS IS A WORK IN PROGRESS - NOT YET IMPLEMENTED

Notes on future replication feature:

  • Master-slave replication
  • Read replicas
  • Automatic failover

TODO: Finish implementation before documenting. =============== END FILE ===============

evals

README.md

tile.json