Dhanesh Purohit
Community@dhanesh · Mumbai, India
I like seeing my name in commit histories | I have involuntary neck and pelvic movements on listening to good music
Agent Skills by Dhanesh Purohit
Showing 33 vetted skills indexed across 3 GitHub repositories.
dhaneshokf-site-kit
Generate a browsable Astro and Starlight static website from an OKF knowledge bundle.
dhaneshmockstar-mock
Generate a runnable mockstar mock server from API specs and documentation.
dhaneshclean-code
Applies Clean Code, Clean Architecture, and Clean Craftsmanship principles to writing, reviewing, and refactoring software.
dhaneshfeynman-walkthrough
Guides understanding-first walkthroughs of codebases, papers, and systems with persistent OKF explainers.
dhaneshai-migration-operating-model
Orchestrates code migrations through rulebooks, parity judges, and phase-gated agent workflows.
dhaneshcontext-hygiene-kit
Installs a bounded scored context cache with lifecycle hooks into Claude Code projects.
dhaneshstarlight-handbook-kit
Scaffold and extend Astro + Starlight documentation handbooks with nine CI integrity gates.
dhaneshverifier-installer
Installs format, build, test commands and a CI workflow into repositories lacking a verify loop.
dhaneshjev-agent-setup
Installs the jev CLI and managed Jev instruction blocks into global agent configuration files.
dhaneshbug-autopsy
Generate blameless, evidence-cited post-mortems for resolved defects and incidents.
dhaneshsecurity-posture-audit
Audit repository security hygiene with deterministic offline checks and severity-graded findings.
dhaneshtmux-agent-herdr-lite
Turns tmux into a cockpit for supervising and coordinating multiple coding agents.
dhaneshbase-in-reality
Audits repository code, architecture, and business logic against fetched academic and standards sources.
dhaneshknowledge-gardener
Detect drift in OKF knowledge bundles and refresh explainers against pinned source fingerprints.
dhaneshrepo2skill
Scaffold and review Agent Skill directories that pass repository quality gates.
dhaneshworld-model-ledger
Installs a SQLite-backed world model that tracks codebase entities, interactions, and constraints with confidence scores.
dhaneshagent-ready-rails
Audits repository readiness for coding agents and scores ten engineering rails from observed evidence.
dhaneshtest-safety-net
Writes proven-failing-first characterization tests into untested Python, Node, Go, and Rust codebases.
dhaneshautoresearch
Automate codebase improvement loops with constraint discovery and baseline capture.
dhaneshmanifold
Surface and manage all constraints before implementation across planning, design, and deployment.
dhaneshmanifold-m0-init
Initialize constraint manifold scaffolds with JSON and Markdown artifacts.
dhaneshmanifold-parallel
Automate parallel task execution with isolated git worktrees and safe merging.
dhaneshmanifold-m3-anchor
Derive required truths from a defined outcome using backward reasoning.
dhaneshmanifold-suggest
Analyze multi-task plans to identify safe parallelization groups and sequential dependencies.
Frequently Asked Questions About Dhanesh Purohit
FAQPage SchemaWhat tasks can I accomplish with Dhanesh Purohit's skills?▼
You can plan features with constraint manifolds and falsifiable specs, run unattended task plans to pull requests, generate mock servers from API specs, author characterization tests for Python/Node/Go/Rust, audit security posture and agent-readiness, publish OKF knowledge bundles as Astro + Starlight sites, and write blameless incident post-mortems.
Who are these skills designed for?▼
Software engineers and engineering teams running coding agents such as Claude Code, Codex, or Gemini CLI. They suit developers handling migrations, legacy codebases without tests, repo hygiene audits, documentation sites, and teams wanting supervised or unattended agent execution with verification gates.
What are the runtime requirements and installation workflow?▼
Most skills need only python3 stdlib and a POSIX shell, running fully offline. Specific skills add requirements: Node.js 18+ for Astro + Starlight sites, Bun or Docker for mockstar mocks, tmux 3.2+ for the agent cockpit, git 2.31+ for worktree-based execution, and Claude Code lifecycle hooks for memory setups.
Are these skills open source and what do they cost?▼
Yes, every skill in the manifest carries an MIT license and is free to use and modify. They are deliberately dependency-light: most run on python3 stdlib alone with no pip installs or network access, keeping adoption cost at zero beyond the host agent environment.
How do the skills handle safety for unattended agent runs?▼
Self-prompting loops bake in hard-stop backstops and trusted/untrusted channel boundaries against prompt injection. The factory-conductor executes only inside explicit autonomy grants, enforces wall-clock and parallel budgets, re-runs verify commands as proof, and leaves PR merging to humans.