@lotagate/cli 0.1.23 → 0.1.25

package/skills/api-contract-review/SKILL.md

@@ -1,17 +0,0 @@
----
-name: api-contract-review
-description: Review an API contract, adapter or protocol flow for compatibility, validation and failure behavior.
-user-invocable: true
-model-invocable: true
-allowed-tools: [filesystem.read, filesystem.list]
----
-# API contract review
-
-Trace the request and response from the public boundary through validation,
-application logic, adapters and mapping.
-
-Check input and output schemas, optionality, backward compatibility, error
-codes, authentication, authorization, timeouts, retries, cancellation,
-pagination, idempotency and sensitive data exposure. Compare the contract with
-callers and tests. Report evidence-backed findings and distinguish runtime
-verification requirements from confirmed behavior.

package/skills/database-review/SKILL.md

@@ -1,17 +0,0 @@
----
-name: database-review
-description: Review database access for correctness, integrity, query behavior and concurrency safety.
-user-invocable: true
-model-invocable: true
-allowed-tools: [filesystem.read, filesystem.list]
----
-# Database review
-
-Trace repositories, queries, transactions and constraints through their
-callers. Inspect filtering, tenant isolation, soft deletion, pagination,
-indexes, joins, read-modify-write races, unique constraints, lock behavior and
-external calls inside transactions.
-
-Do not recommend an index or schema change without query shape, data-volume and
-write-amplification evidence. Label runtime plans and load behavior as needing
-verification when no measurements are available.

package/skills/dependency-review/SKILL.md

@@ -1,16 +0,0 @@
----
-name: dependency-review
-description: Evaluate a dependency change for necessity, security, compatibility and operational cost.
-user-invocable: true
-model-invocable: true
-allowed-tools: [filesystem.read, filesystem.list, shell.exec]
----
-# Dependency review
-
-Check whether an existing abstraction, direct dependency, native API or
-framework capability already solves the problem. If a new package is still
-justified, assess maintenance, license, advisories, transitive dependencies,
-runtime cost, compatibility, migration risk and rollback.
-
-Do not modify manifests or lockfiles unless the user explicitly requests the
-dependency change and the evidence supports it.

package/skills/frontend-review/SKILL.md

@@ -1,18 +0,0 @@
----
-name: frontend-review
-description: Review a frontend flow for state ownership, rendering behavior, accessibility and resilient user states.
-user-invocable: true
-model-invocable: true
-allowed-tools: [filesystem.read, filesystem.list]
----
-# Frontend review
-
-Trace route, page, component, state, request and render behavior. Check state
-ownership, derived state, effect chains, stale responses, cancellation,
-loading/error/empty states, keyboard access, focus, semantics, large lists and
-user-visible failure recovery.
-
-Recommend memoization or component extraction only when the affected subtree,
-trigger and measurable benefit are clear. Preserve the established frontend
-architecture and avoid presentation code becoming the source of truth for
-authorization.

package/skills/incident-response/SKILL.md

@@ -1,18 +0,0 @@
----
-name: incident-response
-description: Investigate an operational failure using a bounded evidence-first response workflow.
-user-invocable: true
-model-invocable: false
-allowed-tools: [filesystem.read, filesystem.list, shell.exec]
----
-# Incident response
-
-Establish the impact, timeline, affected component and current safety boundary
-before changing anything. Preserve evidence and secrets, inspect logs and
-configuration through approved paths, and separate confirmed facts from
-hypotheses.
-
-Prioritize containment, reversible mitigation, data integrity and clear
-rollback. Record commands, observations, decision points and verification. Do
-not delete evidence, rotate credentials, alter production state or apply a
-destructive fix without explicit authorization.

package/skills/performance-review/SKILL.md

@@ -1,17 +0,0 @@
----
-name: performance-review
-description: Find evidence-backed performance bottlenecks and define a measurable safe optimization.
-user-invocable: true
-model-invocable: true
-allowed-tools: [filesystem.read, filesystem.list, shell.exec]
----
-# Performance review
-
-Inspect the complete flow and existing measurements before proposing an
-optimization. Prioritize database and external I/O, duplicate requests,
-blocking CPU work, unbounded concurrency, memory growth and large payloads.
-
-For every finding state the evidence, workload, bottleneck, correctness risks,
-measurement command and acceptance criterion. Do not invent percentage gains or
-recommend caching, memoization, larger pools or infrastructure without an
-invalidation and capacity analysis.

package/skills/refactor/SKILL.md

@@ -1,23 +0,0 @@
----
-name: refactor
-description: Restructure existing code to improve ownership, reuse and testability without changing behavior.
-user-invocable: true
-model-invocable: true
-allowed-tools: [filesystem.read, filesystem.list, filesystem.exists, filesystem.write, shell.exec]
----
-# Refactor
-
-Refactor only when the requested scope has a concrete maintainability or
-correctness benefit.
-
-1. Trace all callers, exports, contracts and tests before moving code.
-2. Separate responsibilities by ownership and dependency direction, not by
-   arbitrary file size.
-3. Reuse existing utilities and preserve public behavior, errors, ordering and
-   side effects.
-4. Keep each extracted unit cohesive and independently testable.
-5. Use a focused diff and verify after each safe extraction, then run the
-   broader regression checks.
-
-Do not combine unrelated cleanup, introduce premature generic abstractions or
-hide behavior changes inside a refactor.

package/skills/release-readiness/SKILL.md

@@ -1,16 +0,0 @@
----
-name: release-readiness
-description: Assess whether a requested change is ready for release across behavior, security, operations and verification.
-user-invocable: true
-model-invocable: false
-allowed-tools: [filesystem.read, filesystem.list, filesystem.exists, shell.exec]
----
-# Release readiness
-
-Inspect the diff, package metadata, build and test configuration, public
-contracts, migrations, configuration, documentation and deployment assumptions.
-
-Check backward compatibility, secret handling, authorization, error and
-rollback behavior, observability, resource cleanup, versioning and file-size
-constraints. Run only documented checks and report blockers, evidence,
-pre-existing failures and unverified runtime risks separately.

package/skills/security-review/SKILL.md

@@ -1,16 +0,0 @@
----
-name: security-review
-description: Review the current workspace for security risks with evidence and safe remediation guidance.
-user-invocable: true
-model-invocable: false
-allowed-tools: [filesystem.exists, filesystem.read, filesystem.list]
----
-
-Review the current workspace for security issues. Trace each relevant input to
-its sensitive sink before reporting it. Prioritize authentication,
-authorization, path traversal, command injection, secret handling, unsafe
-deserialization, and dependency boundaries.
-
-Report only evidence-backed findings. For each finding, include the affected
-file and flow, trigger condition, impact, confidence, and the smallest safe
-remediation. Do not modify files unless the user explicitly asks for a fix.

package/skills/test-plan/SKILL.md

@@ -1,22 +0,0 @@
----
-name: test-plan
-description: Design a focused test plan covering behavior, errors, security and integration boundaries.
-user-invocable: true
-model-invocable: true
-allowed-tools: [filesystem.read, filesystem.list]
----
-# Test plan
-
-Create an implementation-ready test plan for the requested change.
-
-Cover:
-
-- happy paths and important edge cases;
-- runtime validation and stable error behavior;
-- authentication, authorization and trust boundaries;
-- cancellation, timeout and bounded-concurrency behavior;
-- persistence, external adapters and malformed input;
-- the narrow unit test, integration test and final regression command.
-
-For each test, state the setup, action, expected result and why it protects a
-business invariant. Prefer behavior assertions over implementation details.

package/skills/verify/SKILL.md

@@ -1,13 +0,0 @@
----
-name: verify
-description: Identify and run the narrowest relevant verification for the requested workspace change.
-user-invocable: true
-model-invocable: false
-allowed-tools: [filesystem.exists, filesystem.read, filesystem.list, shell.exec]
----
-
-Inspect the workspace documentation and package configuration to determine the
-narrowest relevant verification for the requested scope. Run only commands
-that are documented by the project. Report each command and its result,
-separating pre-existing failures from failures introduced by the requested
-change. Do not edit source files as part of verification.