All authors
Claude Skills by gethamster
github.com/gethamster604 skills52 installs632 views
- Executing With Implicit GuidanceAct from prepared understanding so familiar cases flow from Orient to Act, and treat each action as a test that feeds the next cycle.Votes: 0GitHub stars: 2
- Facilitating Osmotic CommunicationArrange seating, sightlines and team habits so useful information spreads by overhearing, while quiet time and private space stay protected.Votes: 0GitHub stars: 2
- Facilitating Participatory IdeationRun ideation sessions where users and stakeholders generate ideas together, then converge on concepts worth prototyping.Votes: 0GitHub stars: 2
- Formulating Atomic Decision QuestionsTurn each fork in an agent into a typed question with a fixed answer space and a written rubric, so a decision model answers and code branches.Votes: 0GitHub stars: 2
- Formulating Typed Decision QuestionsWrite Choice, Score and Noul questions that each ask one judgment about a state and return a typed answer your code can branch on.Votes: 0GitHub stars: 2
- Framing Human Centered ProblemsTurn empathy research into a point-of-view statement naming a specific user, a real need, and an insight that reframes the challenge.Votes: 0GitHub stars: 2
- Generating Divergent IdeasFacilitate a time-boxed brainstorm that produces many varied ideas against a framed problem, with judgment saved for a separate selection step.Votes: 0GitHub stars: 2
- Handling Abstention And UncertaintySet per-task confidence thresholds from logged outcomes so software acts on reliable decisions and escalates uncertain ones to people.Votes: 0GitHub stars: 2
- Implementing Agent Planning StrategiesPlan multi-step agent work in Semantic Kernel by replacing deprecated planners with an automatic function-calling loop over registered plugins.Votes: 0GitHub stars: 2
- Implementing Devops With Continuous Delivery PipelineBuild and run the four-stage SAFe pipeline that moves small batches from customer insight to on-demand release, with feedback closing the loop.Votes: 0GitHub stars: 2
- Implementing Frequent Delivery CyclesPlan and run delivery cycles that put working, tested, usable software in front of real users on a steady, context-fit rhythm.Votes: 0GitHub stars: 2
- Implementing Internal Linking For Topic ClustersWire hubs, spokes and selective cross-cluster links with descriptive anchors so readers and crawlers can reach every page in a topic cluster.Votes: 0GitHub stars: 2
- Integrating Expert User AccessSecure a real expert user, keep a continuous question channel open, and turn their feedback into requirement, priority and plan changes.Votes: 0GitHub stars: 2
- Integrating External Tools And ApisTurn APIs, queries and code into described, schema-bound LangChain tools that an agent can select, call and recover from safely.Votes: 0GitHub stars: 2
- Integrating Plugins And Tools Into AgentsExpose native code, prompt templates, OpenAPI services and MCP tools as Semantic Kernel plugins an agent can discover and call.Votes: 0GitHub stars: 2
- Iterating Design Solutions With UsersRun repeated build, test and refine cycles with real users, turn feedback into specific revisions, and decide when to pivot, persevere or stop.Votes: 0GitHub stars: 2
- Iterating From EvidenceRead test results, decide which design thinking mode to return to, and update the problem and solution until the evidence supports shipping.Votes: 0GitHub stars: 2
- Launching Agile Release TrainsDefine a value stream, form cross-functional teams and roles, and confirm readiness so a new agile release train can plan and deliver together.Votes: 0GitHub stars: 2
- Layering Instruction Hierarchies In PromptsSort Claude instructions into persistent, on-demand and turn-level layers so each layer does one job and the context stays lean.Votes: 0GitHub stars: 2
- Loading And Splitting DocumentsTurn raw files into well-sized, metadata-rich chunks that an LLM pipeline can embed, retrieve and cite accurately.Votes: 0GitHub stars: 2
- Managing Context Window Token BudgetsKeep Claude Code's context window from filling and degrading by budgeting what enters it and compacting or clearing on purpose.Votes: 0GitHub stars: 2
- Managing Lean Portfolio With SafeGive one portfolio group clear decision rights and move epics through a Portfolio Kanban, committing capacity only as evidence builds.Votes: 0GitHub stars: 2
- Managing Memory And Conversation StateDecide what a LangChain agent keeps, trims, summarizes and persists so long conversations stay coherent and inside the context window.Votes: 0GitHub stars: 2
- Managing Shared Task StateKeep an Agent Team's shared task list accurate so claims, completions and dependencies match reality and blocked work unblocks on time.Votes: 0GitHub stars: 2
- Mapping Customer OpportunitiesSynthesize customer interview insights into a structured map of opportunities, then prioritize which one your team should pursue next.Votes: 0GitHub stars: 2
- Orchestrating Multi Agent ConversationsCoordinate several agents in one workflow with bounded roles, explicit handoffs, stop rules, protocol boundaries and a token budget.Votes: 0GitHub stars: 2
- Parallelizing Independent Work Across SessionsDecide which Agent Team tasks can safely run at once in a shared codebase, remove dependencies with stubs, and weigh the token cost.Votes: 0GitHub stars: 2
- Planning Human Centered ImplementationTurn a tested human-centered concept into a pilot and rollout plan that people want, the organization can deliver, and the business can sustain.Votes: 0GitHub stars: 2
- Planning Program IncrementsRun a PI Planning event that turns business context into team iteration plans, visible dependencies, named risks and committed PI objectives.Votes: 0GitHub stars: 2
- Preventing Context Poisoning And ClashesDiagnose poisoned, noisy, confused or contradictory context in Claude sessions and fix it at its source, not just in the chat.Votes: 0GitHub stars: 2
- Prioritizing With WsjfRank SAFe features and epics by dividing a relative cost of delay by relative job size, so the backlog delivers the most value soonest.Votes: 0GitHub stars: 2
- Providing Individualized ConsiderationCoach each team member according to their own motivations, strengths and goals, and keep every coaching commitment you make.Votes: 0GitHub stars: 2
- Recovering From Verification FailuresDecide what happens after a failed check: feed evidence back, retry within bounds, escalate to a stronger model, or hand off to a human.Votes: 0GitHub stars: 2
- Representing Current Agent StateBuild the state an agent is judged against, shape it as text or structured data, and persist it outside the model so progress survives every run.Votes: 0GitHub stars: 2
- Reviewing And Synthesizing Teammate OutputsCombine teammate outputs into one change, run cross-module checks, resolve conflicts, and accept work only against explicit success criteria.Votes: 0GitHub stars: 2
- Running Assumption TestsIdentify the riskiest assumptions behind a product idea and run small, fast experiments that produce evidence before your team commits to building.Votes: 0GitHub stars: 2
- Running Inspect And Adapt WorkshopsFacilitate the end-of-PI event that demos the real solution, reviews results against PI objectives, and produces owned improvements for the next PI.Votes: 0GitHub stars: 2
- Running Reflection WorkshopsRun a short, regular team workshop that reviews what worked and what did not, then turns the findings into changes the team actually applies.Votes: 0GitHub stars: 2
- Scanning Environment For SignalsScan internal and external sources in Observe, favor leading indicators, and flag weak signals before they reach lagging reports.Votes: 0GitHub stars: 2
- Selecting And Comparing Agent ArchitecturesChoose between ChatCompletionAgent, OpenAIAssistantAgent or a custom agent by profiling needs, testing side by side and recording the decision.Votes: 0GitHub stars: 2
- Selecting And Executing Bounded ActionsTurn a model's typed answer into one validated action from a closed set, gated by confidence, with a no-match route and a human fallback.Votes: 0GitHub stars: 2
- Selecting Crystal Color VariantAssess team size and project criticality, then choose Crystal Clear, Yellow or Orange, and recognise when no Crystal variant fits.Votes: 0GitHub stars: 2
- Separating Generation From Decision MakingAudit an agent run, label every operation as text, decision, or rule, and route each one to an LLM, to Jev, or to deterministic code.Votes: 0GitHub stars: 2
- Shortening Feedback Loop CyclesCut the waiting out of your decision cycle with predefined signposts, if-then rules, standing orientation cells and delegated authority.Votes: 0GitHub stars: 2
- Splitting Features Into StoriesSize epics, features and stories correctly, then split oversized features into iteration-sized user stories and enablers a team can finish.Votes: 0GitHub stars: 2
- Stimulating Intellectual ChallengeGet your team questioning assumptions and testing new approaches through bounded experiments, without losing execution discipline.Votes: 0GitHub stars: 2
- Story Mapping Customer ExperiencesBuild an experience map from customer interview stories to see the full journey and pinpoint the moments where product opportunities live.Votes: 0GitHub stars: 2
- Structuring Machine To Machine Decision OutputsTreat a decision model's typed output and confidence value as a versioned software contract that code validates, routes and acts on.Votes: 0GitHub stars: 2
- Structuring Retrieval Augmented ContextSplit documents by heading, add situating context to each chunk, index it for semantic and keyword search, and hand Claude structured results.Votes: 0GitHub stars: 2
- Structuring Shared Agent StateWrite the state snapshot every agent decision reads, with goal, evidence and gaps, and write each action's real result back before the next call.Votes: 0GitHub stars: 2