AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
taoluo avatar

taoluo

5 listings · 0 installs

Open-source publisher. Listings imported from github.com/taoluo — credited to the original author with their license.

↗ github.com/taoluo
5 results
Self-run
SKILL

Scope Triage

Classifies feature-plan work items into implementation-scope buckets (forbidden, correctness floor, happy path, observability, lightweight defensive, dev-only, release blocker, deferred hardening, polish, overengineering) and produces a Scope Delta Matrix that surfaces where the classification disagrees with the plan's declared scope. The artifact is for human reviewers; agent handoff happens onl…

0
32
Free
Self-run
SKILL

Review Report

Spawns parallel reviewer subagents from a code-review brief, collects their structured shard reports, and aggregates them into a findings-first review report. Use after code-review-plan to actually execute the parallel review and produce a decision-ready report.

0
32
Free
Self-run
SKILL

Debug Log

Records smoke-test and execution-time debugging findings as an append-only, plan-linked bug log. Use during implementation debugging to capture observed behavior, expected behavior, root cause, fix, verification evidence, regression coverage, and plan-revision implications.

0
21
Free
Self-run
SKILL

Code Review Plan

Creates a parallelizable code-review workplan from a source plan. Optional sibling artifacts (TLDR, scope, debug log) and code-diff hints sharpen the workplan but are not required. Use whenever a planned change has reached a state worth reviewing — pre- or post-smoke — to identify review shards, review angles, adversarial prompts, subagent assignments, and final aggregation requirements.

0
32
Free
Self-run
SKILL

Tldr Plan

Compile a raw implementation plan into a compact-first, self-contained TLDR artifact for human audit before agent implementation. Extract problem context, assumptions, scope, hard constraints, D0-D6 decision trace, critical diagrams, evidence, and stop conditions. Iterative — re-run after each plan revision until the human audit passes; only then hand the plan to a coding agent. Use when a plan i…

0
22
Free
You've reached the end · 5 loaded