[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7FDHNRFBVTSPVP0FE5BMDG2":3,"skill-same:01M7FDHNRFBVTSPVP0FE5BMDG2":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":29,"updatedAt":30,"install":31,"resources":44},"01M7FDHNRFBVTSPVP0FE5BMDG2","simulation",null,"Design Monte Carlo simulations for theoretical validation: sample size determination, convergence diagnostics, and result presentation. Specialized for statistical methodology papers.","Abdulqudduz\u002Fclaude-anvil","Abdulqudduz","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F111210054?v=4","01M7FDCB600A6D0CXP9AZ75K5G","MAQGEEK","01M7FDCB6K7XH3DSTYWD1B4GVB",0,"MIT","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fpapermill\u002Fskills\u002Fsimulation\u002FSKILL.md",[],[],"2026-11-16 10:05:34",false,"---\nname: simulation\ndescription: >-\n  Design Monte Carlo simulations for theoretical validation: sample size\n  determination, convergence diagnostics, and result presentation.\n  Specialized for statistical methodology papers.\n---\n\n# Monte Carlo Simulation Design\n\nHelp the researcher design rigorous Monte Carlo simulations to validate theoretical results. Simulations bridge the gap between theory and practice: they demonstrate that analytical formulas work as predicted and reveal finite-sample behavior.\n\n## Step 1: Read Context\n\nRead `.papermill\u002Fstate.md` (Read tool) for:\n- **Thesis**: What theoretical result needs validation.\n- **Experiments**: Any existing simulation experiments.\n\nIf `.papermill\u002Fstate.md` does not exist, ask the user which theoretical result needs validation. Simulation design can proceed without the state file: suggest running `\u002Fpapermill:init` afterward.\n\nScan the repository for existing simulation code and results (Glob\u002FRead tools).\n\n## Step 2: Identify What to Validate\n\nAsk: \"Which theoretical result are you validating with this simulation?\"\n\nCommon simulation targets in methodology papers:\n- **Asymptotic distributions**: Does the CLT approximation hold at finite n?\n- **Estimator properties**: Is the MLE unbiased? What is its MSE?\n- **Covariance\u002Finformation matrices**: Does the empirical covariance match the theoretical Fisher information?\n- **Confidence interval coverage**: Do 95% CIs actually cover 95% of the time?\n- **Power analysis**: Can the test detect a given effect size?\n- **Algorithm convergence**: Does the optimizer find the right answer?\n\n## Step 3: Design the Simulation\n\n### Parameters\n\n- **True parameter values**: Choose several configurations spanning easy, moderate, and hard cases.\n- **Sample sizes**: Include small (n=20-50), moderate (n=100-500), and large (n=1000+) to show convergence.\n- **Number of replications**: Typically R=1,000-10,000. More replications reduce Monte Carlo error but cost time. The Monte Carlo standard error for a proportion p is sqrt(p(1-p)\u002FR).\n\n### Data Generation\n\nSpecify exactly how to generate synthetic data:\n1. Set random seed for reproducibility.\n2. Generate from the known model with known true parameters.\n3. Apply the estimation\u002Finference procedure to each replicate.\n4. Collect the quantities of interest.\n\n### Metrics\n\n| What to measure | How to summarize |\n|----------------|-----------------|\n| Bias | Mean(estimate) - true value |\n| Variance | Var(estimates) across replicates |\n| MSE | Bias^2 + Variance |\n| Coverage | Fraction of CIs containing true value |\n| Convergence rate | Plot metric vs. n on log scale |\n\n### Memory and Compute\n\n- **Batch processing**: If R * n is large, process in batches to control memory. Accumulate sufficient statistics rather than storing all raw data.\n- **Parallelization**: Independent replicates are embarrassingly parallel. Use `parallel` (R), `multiprocessing` (Python), or OpenMP (C++).\n- **Progress reporting**: For long simulations, report progress every ~10% of replicates.\n\n## Step 4: Convergence Diagnostics\n\nBefore presenting results, verify the simulation itself is reliable:\n\n1. **Monte Carlo standard errors**: Report them alongside point estimates. If MCSE is large relative to the quantity of interest, increase R.\n2. **Trace plots**: Plot running averages to verify convergence.\n3. **Sensitivity to seed**: Run with 2-3 different seeds. Results should be stable.\n\n## Step 5: Result Presentation\n\nDesign the output tables and figures:\n\n### Tables\n- Rows: parameter configurations or sample sizes\n- Columns: bias, SD, MSE, coverage, or whatever metrics are relevant\n- Include Monte Carlo standard errors in parentheses\n\n### Figures\n- **Convergence plots**: Metric vs. n (log scale) with theoretical prediction overlaid\n- **QQ plots**: Compare empirical distribution of estimator to theoretical asymptotic distribution\n- **Box plots**: Distribution of estimates across replicates for each configuration\n\n## Step 6: Common Pitfalls\n\nWarn about:\n\n- **Insufficient replicates**: R=100 is rarely enough. For coverage studies, you need R >= 1,000.\n- **Ignoring Monte Carlo error**: Always report MCSE. A coverage of 0.94 with MCSE=0.007 is consistent with 0.95.\n- **Conditioning bias**: If you filter out \"non-convergent\" replicates, the remaining sample is biased. Report the rejection rate.\n- **Single parameter configuration**: Results at one parameter value do not generalize. Test across a range.\n- **Floating-point issues**: For extreme parameter values, numerical issues can corrupt results. Include sanity checks.\n\n## Step 7: Update State File\n\nIf `.papermill\u002Fstate.md` exists, register the simulation (Edit tool) under `experiments`. If it does not exist, skip registration and suggest running `\u002Fpapermill:init`.\n\nThe entry uses the standard experiment schema with an optional `config` block for simulation-specific parameters:\n\n```yaml\nexperiments:\n  - name: \"simulation-name\"\n    type: \"simulation\"\n    hypothesis: \"Empirical covariance matches theoretical FIM as n grows\"\n    status: \"planned\"\n    script: \"research\u002Fsimulate_covariance.R\"\n    last_run: null\n    config:  # optional experiment field (see docs\u002Fstate-schema.md)\n      replications: 5000\n      sample_sizes: [50, 100, 200, 500, 1000]\n      parameter_configs: 3\n```\n\nAppend a timestamped note documenting the simulation design.\n\n## Step 8: Suggest Next Steps\n\nBased on the simulation status, suggest the most relevant next step:\n\n- **Implement**: \"Write the simulation script. Start with a small pilot (R=100) to debug before the full run.\"\n- If the simulation validates a proof → \"Once results confirm the theory, integrate both the proof and simulation evidence into the paper. Use `\u002Fpapermill:proof` if the proof itself needs work.\"\n- If simulation results are surprising or contradict theory → \"The simulation suggests the theoretical result may need revisiting. Use `\u002Fpapermill:proof` to re-examine the proof's assumptions.\"\n- If results are written up → \"Use `\u002Fpapermill:review` to get feedback on the presentation of simulation results.\"\n","# Monte Carlo Simulation Design\n\nHelp the researcher design rigorous Monte Carlo simulations to validate theoretical results. Simulations bridge the gap between theory and practice: they demonstrate that analytical formulas work as predicted and reveal finite-sample behavior.\n\n## Step 1: Read Context\n\nRead `.papermill\u002Fstate.md` (Read tool) for:\n- **Thesis**: What theoretical result needs validation.\n- **Experiments**: Any existing simulation experiments.\n\nIf `.papermill\u002Fstate.md` does not exist, ask the user which theoretical result needs validation. Simulation design can proceed without the state file: suggest running `\u002Fpapermill:init` afterward.\n\nScan the repository for existing simulation code and results (Glob\u002FRead tools).\n\n## Step 2: Identify What to Validate\n\nAsk: \"Which theoretical result are you validating with this simulation?\"\n\nCommon simulation targets in methodology papers:\n- **Asymptotic distributions**: Does the CLT approximation hold at finite n?\n- **Estimator properties**: Is the MLE unbiased? What is its MSE?\n- **Covariance\u002Finformation matrices**: Does the empirical covariance match the theoretical Fisher information?\n- **Confidence interval coverage**: Do 95% CIs actually cover 95% of the time?\n- **Power analysis**: Can the test detect a given effect size?\n- **Algorithm convergence**: Does the optimizer find the right answer?\n\n## Step 3: Design the Simulation\n\n### Parameters\n\n- **True parameter values**: Choose several configurations spanning easy, moderate, and hard cases.\n- **Sample sizes**: Include small (n=20-50), moderate (n=100-500), and large (n=1000+) to show convergence.\n- **Number of replications**: Typically R=1,000-10,000. More replications reduce Monte Carlo error but cost time. The Monte Carlo standard error for a proportion p is sqrt(p(1-p)\u002FR).\n\n### Data Generation\n\nSpecify exactly how to generate synthetic data:\n1. Set random seed for reproducibility.\n2. Generate from the known model with known true parameters.\n3. Apply the estimation\u002Finference procedure to each replicate.\n4. Collect the quantities of interest.\n\n### Metrics\n\n| What to measure | How to summarize |\n|----------------|-----------------|\n| Bias | Mean(estimate) - true value |\n| Variance | Var(estimates) across replicates |\n| MSE | Bias^2 + Variance |\n| Coverage | Fraction of CIs containing true value |\n| Convergence rate | Plot metric vs. n on log scale |\n\n### Memory and Compute\n\n- **Batch processing**: If R * n is large, process in batches to control memory. Accumulate sufficient statistics rather than storing all raw data.\n- **Parallelization**: Independent replicates are embarrassingly parallel. Use `parallel` (R), `multiprocessing` (Python), or OpenMP (C++).\n- **Progress reporting**: For long simulations, report progress every ~10% of replicates.\n\n## Step 4: Convergence Diagnostics\n\nBefore presenting results, verify the simulation itself is reliable:\n\n1. **Monte Carlo standard errors**: Report them alongside point estimates. If MCSE is large relative to the quantity of interest, increase R.\n2. **Trace plots**: Plot running averages to verify convergence.\n3. **Sensitivity to seed**: Run with 2-3 different seeds. Results should be stable.\n\n## Step 5: Result Presentation\n\nDesign the output tables and figures:\n\n### Tables\n- Rows: parameter configurations or sample sizes\n- Columns: bias, SD, MSE, coverage, or whatever metrics are relevant\n- Include Monte Carlo standard errors in parentheses\n\n### Figures\n- **Convergence plots**: Metric vs. n (log scale) with theoretical prediction overlaid\n- **QQ plots**: Compare empirical distribution of estimator to theoretical asymptotic distribution\n- **Box plots**: Distribution of estimates across replicates for each configuration\n\n## Step 6: Common Pitfalls\n\nWarn about:\n\n- **Insufficient replicates**: R=100 is rarely enough. For coverage studies, you need R >= 1,000.\n- **Ignoring Monte Carlo error**: Always report MCSE. A coverage of 0.94 with MCSE=0.007 is consistent with 0.95.\n- **Conditioning bias**: If you filter out \"non-convergent\" replicates, the remaining sample is biased. Report the rejection rate.\n- **Single parameter configuration**: Results at one parameter value do not generalize. Test across a range.\n- **Floating-point issues**: For extreme parameter values, numerical issues can corrupt results. Include sanity checks.\n\n## Step 7: Update State File\n\nIf `.papermill\u002Fstate.md` exists, register the simulation (Edit tool) under `experiments`. If it does not exist, skip registration and suggest running `\u002Fpapermill:init`.\n\nThe entry uses the standard experiment schema with an optional `config` block for simulation-specific parameters:\n\n```yaml\nexperiments:\n  - name: \"simulation-name\"\n    type: \"simulation\"\n    hypothesis: \"Empirical covariance matches theoretical FIM as n grows\"\n    status: \"planned\"\n    script: \"research\u002Fsimulate_covariance.R\"\n    last_run: null\n    config:  # optional experiment field (see docs\u002Fstate-schema.md)\n      replications: 5000\n      sample_sizes: [50, 100, 200, 500, 1000]\n      parameter_configs: 3\n```\n\nAppend a timestamped note documenting the simulation design.\n\n## Step 8: Suggest Next Steps\n\nBased on the simulation status, suggest the most relevant next step:\n\n- **Implement**: \"Write the simulation script. Start with a small pilot (R=100) to debug before the full run.\"\n- If the simulation validates a proof → \"Once results confirm the theory, integrate both the proof and simulation evidence into the paper. Use `\u002Fpapermill:proof` if the proof itself needs work.\"\n- If simulation results are surprising or contradict theory → \"The simulation suggests the theoretical result may need revisiting. Use `\u002Fpapermill:proof` to re-examine the proof's assumptions.\"\n- If results are written up → \"Use `\u002Fpapermill:review` to get feedback on the presentation of simulation results.\"\n",{"name":5,"description":7},"main","f6f3133da90d79585262853f985437619b96f30d","papermill\u002Fskills\u002Fsimulation","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil","https:\u002F\u002Fgithub.com\u002FAbdulqudduz","Hi there! 👋\r\nI'm Muhammed Abdulquddus, a software developer passionate about Flutter and Python, based in Abuja, Nigeria.","2026-10-23 13:31:03",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add Abdulqudduz\u002Fclaude-anvil --skill simulation",{"tool":37,"command":38},"bunx","bunx skills add Abdulqudduz\u002Fclaude-anvil --skill simulation",{"tool":40,"command":41},"pnpm","pnpm dlx skills add Abdulqudduz\u002Fclaude-anvil --skill simulation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fpapermill\u002Fskills\u002Fsimulation\u002FSKILL.md",true,[],[46,53,60,67,74,81],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":19},"01M7FDMZAFJXWE6EQJRZE0RQCA","worldsmith-methodology","Use when working in a project with a worldbuilding documentation ecosystem (docs\u002F or lore\u002F directory). Provides the documentation-first methodology: canonical hierarchy, doc roles (timeline, lore, characters, systems), propagation discipline, and the dual canonical and exploratory workflow.","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fworldsmith\u002Fskills\u002Fworldsmith-methodology\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M7FDMTEXZNE8RXCTCJVXFE4P","prose-craft","Use when writing fiction, editing fiction prose, reviewing chapters, or drafting scenes. Covers prose craft, dialogue mechanics, scene structure, and common failure modes in AI-assisted fiction. Trigger phrases: \"write fiction\", \"edit chapter\", \"prose review\", \"scene draft\", \"dialogue check\", \"fiction craft\", \"edit this scene\", \"review my prose\".","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fworldsmith\u002Fskills\u002Fprose-craft\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M7FDMN0VRXW0G92593BJ5DTS","vista","This skill should be used when the user asks about research directions, open problems, future work, or follow-up research from the academic literature. Trigger phrases include \"open problems in X\", \"what's next for Y\", \"future work for paper Z\", \"research directions on T\", \"salient follow-up research\", \"what should I work on next in\", \"find me follow-up research unrelated to my prior work\", \"broad survey of W\", \"neglected directions in V\". Routes the request to the right Vista MCP tool, reads back the structured paper sections, and synthesizes research directions in the conversation.","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fvista\u002Fskills\u002Fvista\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M7FDMHKJX5F49JEGRPZXWQH9","synthesize","Make a research-agent run conclude and write synthesis.md (outcome, key findings, failed approaches, open questions, recommendations). Signals a live researcher to wrap up, or launches one in synthesis mode when none is running. Use when results are good enough or the agent is stalling.","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fresearch-agent\u002Fskills\u002Fsynthesize\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M7FDMDJT58GYN0YHMWPH3RH5","status","Show the current state of a research-agent run from .research\u002F: liveness, cycles and budget, sub-problem and hypothesis statuses, eval trend, current focus, and branches. Read-only.","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fresearch-agent\u002Fskills\u002Fstatus\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7FDM90XM3H52XNDNKBB5SAP","resume","Resume an interrupted research-agent run from .research\u002F. Re-launches the researcher to reload its structured state and continue from the current focus, optionally on a named branch or with a new budget. Use after a session restart, crash, or explicit pause.","https:\u002F\u002Fgithub.com\u002FAbdulqudduz\u002Fclaude-anvil\u002Fblob\u002Fmain\u002Fresearch-agent\u002Fskills\u002Fresume\u002FSKILL.md",[],[]]