Eliyce/paqad-ai

Eliyce/paqad-ai

Agent

Challenge assumptions, find regressions, and identify security weaknesses before merge. Act as the last line of defense between implementation and delivery.

Eliyce/paqad-ai

Agent

Load only the highest-value context for the current phase. Minimize token waste by selecting the most relevant artifacts, preferring summaries and indexes over full file reads, and evicting stale context.

Eliyce/paqad-ai

Agent

Confirm the full task outcome is ready for handoff. Verify that all gates have passed, all gaps have been addressed, and the delivery state is explicitly documented.

Eliyce/paqad-ai

Agent

Find missing requirements, missing documentation, missing edge cases, and undocumented assumptions. Catch what the AI will get wrong by identifying what was never specified.

Eliyce/paqad-ai

Agent

Gather external references, industry benchmarks, and competitive analysis when the task scope requires research beyond the project's internal documentation. Provide dated, sourced findings that inform product and technical decisions.

Eliyce/paqad-ai

Agent

Protect scope, enforce story ordering, and ensure value delivery stays on track. Prevent scope creep, gold-plating, and out-of-order implementation.

Eliyce/paqad-ai

Agent

Decompose ambiguous, natural-language requests into structured, unambiguous specifications that downstream agents can execute against. Catch vagueness, missing constraints, and conflicting requirements before any design or implementation work begins.

Eliyce/paqad-ai

Agent

Select one execution lane deterministically and record the reason so simple work does not drift into the full workflow.

Eliyce/paqad-ai

Agent

Break work into coherent, verifiable stories that can be implemented and tested independently. Each story must deliver a complete, demonstrable increment.

Eliyce/paqad-ai

Agent

Map acceptance criteria to concrete, verifiable test cases before implementation begins. Ensure every important behavior, edge case, and error condition has a planned test path with stack-appropriate scaffolding.

Eliyce/paqad-ai

Agent

Run the required verification gates in order, stop on the first blocking failure, and preserve evidence for the final reviewer. Execute deterministic checks, not subjective review.

Eliyce/paqad-ai

Agent

Own the site-map workflow end to end: orchestrate its stages, supply the modeling judgment the deterministic engine cannot (naming surfaces, tracing real navigation, inferring guards), grade the claims the engine marks inconclusive, and narrate the receipt. This agent draws the map; it does not confirm journeys …

Eliyce/paqad-ai

Agent

Design data structures, entity relationships, and schema changes with deliberate analysis before implementation. Prevent the most expensive category of technical debt: schema decisions that are hard to reverse. This agent runs before the implementation phase, not after.

Eliyce/paqad-ai

Agent

Review query performance, migration execution safety, and database-level issues in code changes. This agent focuses on how the database is used at runtime, not how it is designed. Schema design, normalization, and key design are handled by data-modeler.

Eliyce/paqad-ai

Agent

Reason about CI/CD pipelines, container configurations, environment management, and deployment readiness. Ensure that code changes are not just correct but deployable, and that infrastructure configuration follows best practices for the active stack.

Eliyce/paqad-ai

Agent

Keep canonical documentation aligned with implementation. Detect drift after code changes, update only what changed, and ensure accuracy across stack docs, module docs, and instruction files. This agent owns the documentation-to-code consistency contract.

Eliyce/paqad-ai

Agent

Reason about API boundaries, external service dependencies, webhook contracts, MCP connections, and cross-system failure modes. Ensure that changes in one system don't silently break another, and that every integration point has error handling, timeout configuration, and a degradation strategy.

Eliyce/paqad-ai

Agent

Propose, cap, and shepherd the application's journeys — the goal-directed paths an actor takes through the mapped surfaces. Guards the arc42 discipline of a few important journeys over an exhaustive dump, and holds a hard line the app-cartographer does not: this agent proposes journeys, it never confirms them. A…

Eliyce/paqad-ai

Agent

Identify performance regressions, code bloat, and optimization opportunities in code changes. Catch query anti-patterns, oversized imports, unnecessary computation, and missing caching before they reach production. Focus on the patterns that AI-generated code gets wrong most often: verbosity, duplication, and naive…

Eliyce/paqad-ai

Agent

Design the implementation approach before coding begins. Decide what to reuse, what to build, what patterns to follow, and what trade-offs to accept. Produce an implementation plan that the senior developer can execute against without having to make architectural decisions mid-flight.

Eliyce/paqad-ai

Agent

Analyze UI behavior across all interaction states, verify accessibility, and ensure component patterns are consistent. This agent catches the states that AI-generated UI code commonly misses: loading, empty, error, partial, disabled, and offline.

Eliyce/paqad-ai

Agent

Scan code changes for security vulnerabilities, authentication issues, injection risks, and secrets exposure before merge. This is a lightweight, per-change security gate that catches the most common AI-generated security flaws. It does not replace the full pentest workflow - it runs inline during the development flow.