Hugefiver
Community@hugefiver · China
I'm a mochine，是时候摸了。
Agent Skills by Hugefiver
Showing 22 vetted skills indexed across 1 GitHub repositories.
hugefiverusing-git-worktrees
Creates isolated git worktrees with project setup and baseline verification before feature development.
hugefiverdeepwork-requesting-code-review
Requests independent code review of implemented changes against goals, criteria, and evidence.
hugefiverast-grep
Search and rewrite code across 25 languages using AST-aware structural patterns.
hugefiverfrontend
Routes frontend UI work through design-system, taste, and Lighthouse audit reference rulesets.
hugefiverinit-deep
Generates hierarchical AGENTS.md knowledge base files across complexity-scored project directories.
hugefiverdeepwork-subagent-driven-development
Coordinates focused subagent workers to execute approved implementation plans with evidence-based verification.
hugefivercoding-agent-sessions
Search and inspect local coding-agent session transcripts across multiple agent platforms.
hugefiverdebugging
Diagnose runtime bugs across languages and binaries with a hypothesis-driven investigation workflow.
hugefiverdeepwork-brainstorming
Clarifies goals, constraints, and risks before implementation through proportional discovery and design.
hugefiverdeepwork-writing-plans
Writes outcome-oriented implementation plans from approved specs before code changes begin.
hugefiverpublish
Verifies ocmm release completion through a fail-closed checker across tags, workflow runs, and release assets.
hugefivergit-master
Performs atomic Git commits, history rewrites, and evidence-based history investigation.
hugefiverremove-ai-slops
Removes AI-generated code slop from branch changes using test-locked parallel cleanup.
hugefiverdeepwork-receiving-code-review
Evaluates code review feedback against evidence before implementing corrections.
hugefiverdeepwork
Routes planning, delegation, and code review tasks through Deepwork workflow semantics inside Codex.
hugefiverdeepwork-dispatching-parallel-agents
Dispatch independent tasks to parallel worker agents and integrate their results.
hugefiverdispatching-parallel-agents
Coordinates independent parallel agent tasks with scoped assignments and merged-result verification.
hugefiverbrainstorming
Produces an approved design before implementing features or behavior changes in dsmm deepwork mode.
hugefiverwriting-plans
Writes structured implementation plans for multi-step work in dsmm deepwork mode.
hugefiverrequesting-code-review
Collects diffs, tests, and verification evidence to request final implementation acceptance review.
hugefiverreceiving-code-review
Verifies reviewer feedback against code before applying fixes.
hugefiversubagent-driven-development
Executes approved implementation plans task-by-task with bounded subagent delegation and review gates.
Frequently Asked Questions About Hugefiver
FAQPage SchemaWhat tasks can I accomplish using Hugefiver's skills?▼
You can manage git worktrees, atomic commits, rebase, bisect, and blame; debug crashes, memory leaks, and flaky tests across languages; run ast-grep codemods; audit frontend performance to Lighthouse 100 via Playwright; remove AI-generated code slop; and coordinate plan-driven, review-gated multi-agent development.
Who are Hugefiver's skills designed for?▼
Software engineers and coding-agent users working in Codex or dsmm deepwork environments. They suit developers needing rigorous git history investigation, cross-language runtime debugging, deterministic syntax-tree rewrites, frontend design and performance audits, and disciplined plan-review-execute cycles with subagent delegation.
How do the deepwork planning and review skills operate in practice?▼
Work starts with brainstorming to clarify outcomes and constraints, then a written multi-step plan. Independent tasks dispatch to parallel subagents, implementation proceeds task-by-task with review gates, and code review is requested and received with verification of feedback before any changes are applied.
What prerequisites or dependencies do these skills require?▼
Skills rely on their referenced runtimes: git for worktree and history operations, ast-grep (sg) for syntax-tree matching, Playwright Chromium for Lighthouse audits, and debuggers such as gdb, lldb, pdb, or delve. Several skills are scoped to dsmm deepwork mode or Codex sessions.
When should the ast-grep skill be used instead of regular search?▼
Use ast-grep only when a task requires exact syntax-tree matching or a deterministic codemod that ripgrep or LSP lookup cannot express reliably, such as rewriting a specific call shape, stripping 'as any', or finding empty catch blocks. Ordinary text search and symbol navigation should not trigger it.