Attribution and Provenance Matrix

Original engineering by Mehansh Barthwal. Officially merged into the community repository alirezarezvani/claude-skills under Pull Request #870.

Enforces verified reasoning before any code is generated.

Original engineering by Mehansh Barthwal. Contributed to alirezarezvani/claude-skills.

Dynamic scraping strategy router with token budgeting and checkpointing.

Original methodology developed and verified during industry research at SVS Aqua Technologies LLP.

Authentic customer complaint extraction with mechanical deduplication.

Original iterative workflow protocol authored by Mehansh Barthwal.

Autonomous self evaluation loop rewriting drafts until rubric bars clear.

Original career engineering skill authored by Mehansh Barthwal.

Extracts operational role requirements into tailored application materials.

Original recruitment audit skill authored by Mehansh Barthwal.

Four stage evaluation gauntlet reflecting actual hiring pipelines.

Original scraping playbook authored by Mehansh Barthwal.

Methodology playbook for mapping urban food delivery markets.

Original notebook manipulation tool authored by Mehansh Barthwal.

Programmatic cell manipulation preserving notebook JSON integrity.

Original routing architecture by Mehansh Barthwal. Routes across AnyDoc (Firecrawl) and MarkItDown (Microsoft).

Intelligent format routing across AnyDoc and MarkItDown.

Examines why large skill collections require machine readable schemas, how semantic token overlap prevents redundant skills, and why automated categorization requires manual supervision.

Why this collection maintains a machine readable registry.

Technical notes exploring how skill files translate across Claude Code, Cursor IDE, Google Antigravity, ChatGPT Custom GPTs, OpenClaw, and local models via Ollama.

How one skill format operates across six agent runtimes.

Explores why traditional prompts fail in production workflows, and how explicit trigger declarations, negative constraints, and output rubrics transform prompts into reliable skills.

The essential difference lies in triggers and negative constraints.

Analyzes how Stop Slop, Humanizer, and Taste Skill tackle the same fundamental challenge: counteracting generic model defaults across text and frontend engineering.

Applying intentional de-biasing to prose and user interfaces.

Agent skill implementation crafted by Mehansh Barthwal based on the original concept and repository by qosi org at https://github.com/qosi-org/arxivist.

Translates scientific research papers into runnable code repositories.

Adapted from Siqi Chen's open source humanizer skill (blader/humanizer). Distributed under the MIT License.

Strips robotic hallmarks and promotional padding from machine prose.

Adapted from Hardik Pandya's open source stop-slop project (hardikpandya/stop-slop). Distributed under the MIT License.

Compact prompt rules that block formulaic filler during initial generation.

Synthesizes technical Unicode cleaning with rules adapted from Stop Slop (Hardik Pandya) and Humanizer (Siqi Chen).

Combines zero width character scrubbing with stylistic rephrasing.

Adapted from DietrichGebert's open source ponytail project. Distributed under the MIT License.

Senior developer minimalism prioritizing standard libraries and YAGNI.

Adapted from Safi Shamsi's open source graphify project. Distributed under Apache 2.0 and MIT licenses.

Constructs queryable knowledge graphs from codebases and documents.

Adapted from the open source Open GSD project. Distributed under the MIT License.

Phase based autonomous development with user acceptance gates.

Adapted from Agent Eyes' open source agent-reach project. Distributed under the MIT License.

Unified research interface and Model Context Protocol across fifteen networks.

Adapted from Bradley Bonanno's open source agent-video project. Distributed under the MIT License.

Transforms video recordings into structured frames and aligned transcripts.

Adapted from Zara Zhang's open source frontend-slides project. Distributed under the MIT License.

Builds animation rich HTML slide decks or converts from PowerPoint.

Open source software authored by Microsoft Corporation. Distributed under the MIT License.

Microsoft utility converting diverse file formats into Markdown.

Adapted from snarktank's open source ralph project. Distributed under the MIT License.

Converts requirements specifications into deterministic agent code.

Adapted from Microsoft Corporation's SkillOpt project. Distributed under the MIT License.

Offline agent review cycle consolidating recurring session corrections.

Protocol implementation inspired by Andrej Karpathy's LLM Council methodology.

Five advisor debate protocol peer reviewing proposals anonymously.

Distilled from public reflections and engineering commentary by Andrej Karpathy.

Four always on instincts preventing common agent coding failures.

Modular design system inspired by contemporary anti slop frontend engineering and typography practices.

Anti slop frontend design direction for web applications.

Rooted in epistemic calibration research and structured intelligence analysis standards.

Strictly separates verified facts from inferences and assumptions.

Architecture context document for tinyhumansai/openhuman.

Architecture context for the tinyhumansai/openhuman codebase.

Personal automation topology and connection state documentation.

Reference configuration for Zapier Model Context Protocol integrations.

Curated engineering runbook compiled across deployments and production workflows.

Engineering playbook and deployment patterns to avoid repeating past errors.