๐ About
I operate at the intersection of Low-Resource Language NLP and Field-Level Community Development. By day, I design data-driven village projects for Tata Steel UISL through TRCSC. By night, I build AI pipelines, Bengali NLP models, and automation tools for underserved communities.
๐ Today I Built
Daily solution repos โ one real problem, one minimal tool, every day.
2026-08-29
stubcheck
Detect silent no-op function stubs (pass/ellipsis bodies) that ship to production and return None โ the unimplemented logic AI leaves behind in callbacks, handlers, and overrides
๐ค Code Quality for AI
๐ Guide
2026-08-28
assertcheck
Detect assert statements used for runtime validation โ catch code that silently vanishes under python -O before it ships to production
๐ Code Quality for AI
๐ Guide
2026-08-20
globalcheck
Detect silent mutation of module-level state in Python โ catch functions that `results.append(...)` to shared globals instead of returning values, the shared-state leak behind untestable AI code
โ
Code Quality for AI
๐ Guide
2026-08-19
broadcheck
Detect bare, broad, and empty except handlers in Python โ catch `except Exception: pass` silent error-swallowing before it masks production crashes
โ
Code Quality for AI
๐ Guide
2026-08-18
mutablecheck
Detect mutable default arguments โ catch `def f(items=[])` shared-state aliasing where state leaks silently between calls, a signature AI-generated-code footgun
โ
Code Quality for AI
๐ Guide
2026-08-16
shadowcheck
Detect identifiers that shadow Python builtins โ catch `list = []` and `def id()` before runtime TypeErrors far from the binding site, an AI-generated-code failure class
โ
Code Quality for AI
๐ Guide
2026-08-15
cccheck
Flag high-cyclomatic-complexity functions in Python โ catch God functions and untestable spaghetti before AI-generated code ships, with McCabe scoring and CI gates
โ
Code Quality for AI
๐ Guide
2026-08-14
dupcheck
Detect duplicated code blocks across a codebase โ catch copy-paste and AI-generated duplicates before they become maintenance debt
โ
Code Quality for AI
๐ Guide
2026-08-13
todolint
Scan codebases for TODO/FIXME/HACK/XXX/BUG markers โ catch unfinished work before it ships, with CI-ready blocker detection and JSON output
โ
Code Quality for AI
๐ Guide
2026-08-12
chunkviz
Visualize recursive character text splitting โ see chunk boundaries, overlap regions, and size distribution before you tune your RAG pipeline
๐ Document Processing
๐ Guide
2026-08-11
fencecheck
Detect unclosed Markdown code fences (``` and ~~~) โ catch broken code blocks that silently swallow the rest of your document
๐ Document Processing
๐ Guide
2026-08-10
mcpconfig
Validate MCP client configuration files (.mcp.json, claude_desktop_config.json, opencode mcp: blocks) โ catch the structural mistakes that silently break MCP servers
๐ค AI Agent Skills/MCP
๐ Guide
2026-08-09
splitcheck
Detect data leakage across ML dataset split files โ flag shared rows before they silently inflate your evaluation metrics
๐ท๏ธ Data Annotation
๐ Guide
2026-08-08
skillcheck
Validate Agent Skills (SKILL.md) directories โ catch missing fields, malformed frontmatter, and broken file references before your agent silently loses a skill
๐ค AI Agent Skills
๐ Guide
2026-08-03
wikicheck
Validate Obsidian-style [[wikilinks]] โ catch broken targets, missing heading anchors, malformed links, and orphan notes in one command
๐ Document Processing
๐ Guide
2026-07-28
linkrot
Detect broken internal Markdown links โ scan your docs for dead file references and missing section anchors
๐ Document Processing
๐ Guide
2026-07-24
labcheck
Validate text annotation datasets (CSV/JSONL) for label conflicts, empty values, and quality issues โ one command, zero config
๐ท๏ธ Data Annotation
๐ Guide
2026-07-23
mcpschema
Validate MCP (Model Context Protocol) tool JSON Schema definitions โ catch missing fields, invalid types, and silent failures before your AI agent calls them
๐ค AI Agent Skills/MCP
๐ Guide
2026-07-23
docstruct
Validate Markdown heading structure โ detects skipped levels, missing H1s, and structural inconsistencies in documentation
๐ Document Processing
๐ Guide
2026-07-22
coco-summary
Zero-dependency CLI to inspect COCO annotation JSON files โ class distribution, bbox stats, and quality flags
๐ท๏ธ Data Annotation
๐ Guide
2026-07-21
tokensnap
Estimate LLM API costs from your terminal โ count tokens, compare 12+ providers
โก CLI Productivity
๐ Guide
2026-07-20
cocostat
COCO dataset statistics from your terminal โ inspect annotation files: class distribution, bbox stats, annotation density
๐ท๏ธ Data Annotation
๐ Guide
2026-07-19
annocheck
Validate YOLO-format CV datasets for corrupted labels, out-of-bounds boxes, missing images, and class mismatches
๐ท๏ธ Data Annotation
๐ Guide
๐ฌ Featured Projects
๐ฎ๐ณ Kalki
India's Open Intelligence Platform โ open-source early-warning signals from public data to foresee climate, health, and economic crises.
View on GitHub โ๐ก๏ธ PromptShield
Defensive LLM structured extraction validation middleware โ guards against prompt injections and repairs truncated JSON.
View on GitHub โ๐ง Obsidian LLM Wiki Kit
LLM-powered knowledge base toolkit for Obsidian โ build and query interlinked wikis with AI-assisted retrieval and synthesis.
View on GitHub โ๐ Bengali Annotation Quality
Bengali annotation quality assessment โ calibrates gold-standard sentiment data with Cohen's Kappa and disagreement analysis.
View on GitHub โ๐ค SW เคชเคฐเคฟเคตเคพเคฐ
Indian Social Work Knowledge Base โ laws, schemes, judgments, constitution & frameworks. Offline-first, zero-install HTML app for field reference.
View on GitHub โโก OpenCode Research Agents
Multi-agent research orchestration inside OpenCode โ parallel epistemic agent pipelines for systematic knowledge discovery.
View on GitHub โ