All authors
Claude Skills by phbui
github.com/phbui46 skills0 installs23 views
- AuditFinal quality gate for pitch implementations. Orchestrator that spawns specialist reviewers (Architecture, Code Quality, Performance, Requirements) then implements fixes in ordered batches. Use after /implement or /build is complete.Votes: 0GitHub stars: 2
- BuildGeneral-purpose builder that implements features, components, or prototypes from any input — a description, spec, design doc, or existing SDD/ADR. Spawns parallel builder agents for independent workstreams. Use to build a feature straight from a design or spec.Votes: 0GitHub stars: 2
- CaptureDeterministically save an idea, preference, or commitment to the gestalt capture inbox — no judgment call, always writes. Use for 'remember this', 'note that', 'I'll do X later', or /capture <text>.Votes: 0GitHub stars: 2
- CavemanUltra-compressed communication mode that slashes token usage ~75% while keeping full technical accuracy.Votes: 0GitHub stars: 2
- CommitCreate a detailed, well-structured commit with interactive review, analysis, and optional push.Votes: 0GitHub stars: 2
- Compact ResumePrepare the session for context compaction or handoff by securing uncommitted work and in-flight processes, then writing one canonical, self-contained resume block to gestalt so a fresh context can resume with zero loss. Use before /compact or a session handoff on a long-running workstream.Votes: 0GitHub stars: 2
- ConsolidateMemory-layer operations: consolidate (prune/merge/promote/cross-link across all four layers), status (health dashboard), recall (query the temporal graph for cross-session history), publish (condensed briefing for a shared Claude Project).Votes: 0GitHub stars: 2
- DiscussCritical rubber-duck session with parallel evidence-gathering agents. Challenges assumptions, steelmans alternatives, pokes holes in a plan before you commit to it, and helps decide between two architectures. Grounds critique in codebase evidence and external research.Votes: 0GitHub stars: 2
- FixSpin up parallel agents to execute a batch of code fixes efficiently. Accepts fixes from linter errors, code review comments, issue lists, descriptions, or TODOs; triages, partitions by file ownership, and verifies results. Use for multiple fixes, errors, or changes across files.Votes: 0GitHub stars: 2
- Help GestaltDisplay a complete guide to all gestalt commands and skills — what each does, when to use it, how they chain together. Use when asked what you can do or what commands/tools are available.Votes: 0GitHub stars: 2
- InvestigateDeep investigation or conceptual deep-dive using codebase and external research. Spawns parallel agents across code, gestalt, docs, and web; synthesizes findings with citations. Use to trace why something breaks, for bugs, technical decisions, or understanding how something works.Votes: 0GitHub stars: 2
- LearnDeep-scan a repository or topic and create/update gestalt knowledge entries. Spawns 4 parallel scanner agents (Infra, Code, Test/Docs, External) for maximum throughput. Use when you need to learn about a repo or create gestalt coverage.Votes: 0GitHub stars: 2
- MutateCreate, modify, or delete gestalt artifacts (commands, skills, rules, agents, references) — e.g. add a new slash command to the system. Handles dual-agent sync, doc updates, and symlink verification automatically.Votes: 0GitHub stars: 2
- PrMulti-persona deep review of a pull request with specialist agents, centralized triage, and selectable fix options.Votes: 0GitHub stars: 2
- PromptOptimize a raw, vague, or underspecified request into a well-structured proper plan that leverages gestalt's full capabilities, then execute it after user confirmation.Votes: 0GitHub stars: 2
- RequirementsRequirements engineering: explain concepts, inspect docs, author pitch/system/SRS/ADR/SDD docs (write down why you chose one approach over another), promote pitch to system, or run implement from spec to code. Subcommands: explain, pitch, system, srs, adr, sdd, promote, implement.Votes: 0GitHub stars: 2
- ResearchDeep online research on a topic or best-practices question, with parallel researcher agents, cross-referenced citations, and synthesized briefings.Votes: 0GitHub stars: 2
- ReviewVerify the gestalt knowledge base against the actual codebase using parallel verifier agents — check that your notes still match the code. Checks every factual claim, validates links and block IDs, finds gaps.Votes: 0GitHub stars: 2
- SaveExtract knowledge from the current conversation and write it to gestalt.Votes: 0GitHub stars: 2
- Swarm CheckCross-repo ripple detection. When a change in one repo may affect others, uses GRAPH.md dependency map to spawn parallel agents checking each affected repo for broken contracts or required updates. Use when making changes that cross service boundaries.Votes: 0GitHub stars: 2
- SyncPull knowledge from external sources (Notion, Linear, Slack) via MCP tools and reconcile with existing gestalt entries. Spawns parallel syncer agents per source. Use to update gestalt with the latest external context.Votes: 0GitHub stars: 2
- VetEval-loop review of a research document (paper, application, dissertation chapter): parallel dimension auditors + venue persona panel, fix-verify rounds to convergence, then cold-referee and verification exit gates.Votes: 0GitHub stars: 2
- Weekly ReviewGTD-style weekly review: triage the capture inbox, resolve/rollover commitments, surface stuck pending items, stale entries, and un-actioned NEEDS APPROVAL lines from recent sessions. Interactive by default, headless summary on a timer.Votes: 0GitHub stars: 2
- AuditFinal quality gate for pitch implementations. Orchestrator that spawns specialist reviewers (Architecture, Code Quality, Performance, Requirements) then implements fixes in ordered batches. Use after /implement or /build is complete.Votes: 0GitHub stars: 2
- BuildGeneral-purpose builder that implements features, components, or prototypes from any input — a description, spec, design doc, or existing SDD/ADR. Spawns parallel builder agents for independent workstreams. Use to build a feature straight from a design or spec.Votes: 0GitHub stars: 2
- CaptureDeterministically save an idea, preference, or commitment to the gestalt capture inbox — no judgment call, always writes. Use for 'remember this', 'note that', 'I'll do X later', or /capture <text>.Votes: 0GitHub stars: 2
- CavemanUltra-compressed communication mode that slashes token usage ~75% while keeping full technical accuracy.Votes: 0GitHub stars: 2
- CommitCreate a detailed, well-structured commit with interactive review, analysis, and optional push.Votes: 0GitHub stars: 2
- Compact ResumePrepare the session for context compaction or handoff by securing uncommitted work and in-flight processes, then writing one canonical, self-contained resume block to gestalt so a fresh context can resume with zero loss. Use before /compact or a session handoff on a long-running workstream.Votes: 0GitHub stars: 2
- ConsolidateMemory-layer operations: consolidate (prune/merge/promote/cross-link across all four layers), status (health dashboard), recall (query the temporal graph for cross-session history), publish (condensed briefing for a shared Claude Project).Votes: 0GitHub stars: 2
- DiscussCritical rubber-duck session with parallel evidence-gathering agents. Challenges assumptions, steelmans alternatives, pokes holes in a plan before you commit to it, and helps decide between two architectures. Grounds critique in codebase evidence and external research.Votes: 0GitHub stars: 2
- FixSpin up parallel agents to execute a batch of code fixes efficiently. Accepts fixes from linter errors, code review comments, issue lists, descriptions, or TODOs; triages, partitions by file ownership, and verifies results. Use for multiple fixes, errors, or changes across files.Votes: 0GitHub stars: 2
- Help GestaltDisplay a complete guide to all gestalt commands and skills — what each does, when to use it, how they chain together. Use when asked what you can do or what commands/tools are available.Votes: 0GitHub stars: 2
- InvestigateDeep investigation or conceptual deep-dive using codebase and external research. Spawns parallel agents across code, gestalt, docs, and web; synthesizes findings with citations. Use to trace why something breaks, for bugs, technical decisions, or understanding how something works.Votes: 0GitHub stars: 2
- LearnDeep-scan a repository or topic and create/update gestalt knowledge entries. Spawns 4 parallel scanner agents (Infra, Code, Test/Docs, External) for maximum throughput. Use when you need to learn about a repo or create gestalt coverage.Votes: 0GitHub stars: 2
- MutateCreate, modify, or delete gestalt artifacts (commands, skills, rules, agents, references) — e.g. add a new slash command to the system. Handles dual-agent sync, doc updates, and symlink verification automatically.Votes: 0GitHub stars: 2
- PrMulti-persona deep review of a pull request with specialist agents, centralized triage, and selectable fix options.Votes: 0GitHub stars: 2
- PromptOptimize a raw, vague, or underspecified request into a well-structured proper plan that leverages gestalt's full capabilities, then execute it after user confirmation.Votes: 0GitHub stars: 2
- RequirementsRequirements engineering: explain concepts, inspect docs, author pitch/system/SRS/ADR/SDD docs (write down why you chose one approach over another), promote pitch to system, or run implement from spec to code. Subcommands: explain, pitch, system, srs, adr, sdd, promote, implement.Votes: 0GitHub stars: 2
- ResearchDeep online research on a topic or best-practices question, with parallel researcher agents, cross-referenced citations, and synthesized briefings.Votes: 0GitHub stars: 2
- ReviewVerify the gestalt knowledge base against the actual codebase using parallel verifier agents — check that your notes still match the code. Checks every factual claim, validates links and block IDs, finds gaps.Votes: 0GitHub stars: 2
- SaveExtract knowledge from the current conversation and write it to gestalt.Votes: 0GitHub stars: 2
- Swarm CheckCross-repo ripple detection. When a change in one repo may affect others, uses GRAPH.md dependency map to spawn parallel agents checking each affected repo for broken contracts or required updates. Use when making changes that cross service boundaries.Votes: 0GitHub stars: 2
- SyncPull knowledge from external sources (Notion, Linear, Slack) via MCP tools and reconcile with existing gestalt entries. Spawns parallel syncer agents per source. Use to update gestalt with the latest external context.Votes: 0GitHub stars: 2
- VetEval-loop review of a research document (paper, application, dissertation chapter): parallel dimension auditors + venue persona panel, fix-verify rounds to convergence, then cold-referee and verification exit gates.Votes: 0GitHub stars: 2
- Weekly ReviewGTD-style weekly review: triage the capture inbox, resolve/rollover commitments, surface stuck pending items, stale entries, and un-actioned NEEDS APPROVAL lines from recent sessions. Interactive by default, headless summary on a timer.Votes: 0GitHub stars: 2