krishamaze

@krishamaze source repo

21 skills
0 followers
7 hours ago last updated

▌
Odoo · krishamaze bundle
Authoritative Odoo 19.0 development knowledge. Use when the user mentions Odoo, odoo module, __manifest__, ORM, account.move, res.partner, ir.model, odoo addon, odoo model, odoo field, odoo view, odoo controller, odoo migration, odoo upgrade, or any Odoo development task. Triggers on: "odoo", "odoo module", "odoo model", "odoo ORM", "odoo 19", "odoo addon", "odoo development", "odoo accounting", "odoo customization". Source: odoo/odoo@19.0 on GitHub. Zero hallucination — all patterns verified from source code and official documentation.

0
installs
▌
Codex MCP · krishamaze bundle
Connect Codex CLI as an MCP server — giving you codex_run and codex_review as native tool calls instead of black-box bash commands. codex_run covers six modes: explore (broad codebase discovery), inspect (targeted read-only and injected-context follow-up), build (write/edit code), debug (reproduce→diagnose→fix→verify), test (write/run tests), research (web search only). codex_review runs independent code review in an isolated thread. Each mode bakes in task-specific instructions so Codex performs well per task type. Use this skill whenever the user mentions: "set up codex MCP", "connect codex to claude", "codex MCP server", "install codex tools", "configure codex integration", or wants Codex available as native tools in any agent. Distributed via `npx skills add` — no global install needed.

0
installs
▌
Fastapi 2026 · krishamaze
FastAPI async Python API patterns for 2026. Use this skill when building API layers for automation agents — route definitions, async handlers, background tasks, WebSocket, Pydantic v2 models, lifespan events, CORS, and production deployment with Gunicorn + Uvicorn workers. ALWAYS use async def for I/O-bound routes. NEVER use sync DB calls in async routes.

0
installs
▌
Camoufox 2026 · krishamaze
Anti-detect stealth browser automation using Camoufox (Firefox-based, C++-level fingerprint spoofing). Use this skill for ANY browser automation that must avoid bot detection — social media logins, persistent sessions, humanized interaction, or scraping protected sites. Covers 2026-current Camoufox API, known instability warnings, async/sync patterns, persistent context, and humanization config. ALWAYS use this over raw Playwright when stealth is required.

0
installs
▌
Project Memory · krishamaze bundle
Maintains DECISIONS.md, CONTEXT.md, STACK.md, and a flat `memory/` knowledge graph. Handles agent handoffs by scanning agent state, consolidating unfinished work, updating memory, and resuming the next action. Also keeps agent config files lean by extracting dumps into `memory/` and maintaining `AGENT.md` symlinks. Invoke when the user says "memory", "save context", "log this decision", "update project docs", "/memory", "/save", "/handoff", "/continue", "resume from another agent", or "what was the last agent doing", or after architectural decisions, resolved failures, completed features, version changes, agent switches, crashes, or rate limits. Trigger aggressively whenever a decision, failure, state change, or handoff should be captured.

0
installs
▌
Socialos · krishamaze bundle
Project memory for SocialOS — unified social inbox with AI reply suggestions. Read at every session start. Trigger when starting any coding task, before using any external API, after any failure, after any architectural decision. Never skip on a new session.

0
installs
▌
Uv Python 2026 · krishamaze
uv — the 2026 standard Python package manager (replaces pip, venv, pyenv, pip-tools, pipx). Use this skill for ALL Python dependency management, virtual environments, Python version management, and Docker Python setups. NEVER use bare pip for new projects in 2026. Covers uv init, uv add, uv sync, lockfiles, pyproject.toml, Docker integration, and migration from pip/Poetry.

0
installs
▌
Devops Handbook · krishamaze bundle
DevOps coaching skill grounded exclusively in The DevOps Handbook, 2nd Edition (Kim, Humble, Debois, Willis, Forsgren). Use this skill whenever a user asks about DevOps practices, CI/CD pipelines, deployment strategies, org design, team topology, value stream mapping, telemetry/observability, DevSecOps, change management, or cultural transformation. Also trigger for questions phrased as "how do I improve my pipeline", "our teams are siloed", "we keep having production incidents", "how do I shift security left", or anything about flow, feedback, and continual learning. This skill acts as a DevOps Coach — it diagnoses problems, guides architectural decisions, and prescribes concrete next steps, always anchored to the book's principles and case studies. Trigger even for general software delivery questions if the DevOps Handbook would have a relevant answer.

0
installs
▌
Docker Vps 2026 · krishamaze
Docker best practices for VPS deployments in 2026. Use this skill when writing Dockerfiles, docker-compose.yml, or managing containers on a Linux VPS. Covers multi-stage builds, uv-based Python images, Xvfb/virtual display containers, volume management, security hardening, and service orchestration with Docker Compose v2. NEVER use legacy docker-compose v1 (hyphenated), old Python base images, or root containers.

0
installs
▌
Gemini API 2026 · krishamaze bundle
Complete Google Gemini API reference for 2026. Use whenever writing code that calls Gemini models. Covers the google-genai SDK, Gemini 3/3.1 models, thought signatures, thinking config, Interactions API, File Search (managed RAG), Computer Use, URL Context, Nano Banana image gen, Live API, ephemeral tokens, TTS, Veo video gen, Lyria music gen, and all tools. ALWAYS prefer `from google import genai` over any legacy import. Use this skill for ANY Gemini API question, even simple ones.

0
installs
▌
Google Flow Labs · krishamaze
Expert knowledge of Google Flow (flow.google) — Google Labs' unified AI creative studio for image and video generation as of 2026. Use this skill whenever a user mentions Flow, Google Flow, Nano Banana, Veo 3.1, Whisk, ImageFX, Google Labs FX, AI filmmaking with Google tools, or asks how to generate/edit images or videos using Google's AI stack. Also trigger when users ask about Google AI Pro/Ultra subscriptions in the context of creative generation, SynthID watermarks, Ingredients to Video, Frames to Video, or asset management in Flow. Use this skill even if the user only tangentially references these tools, e.g. "I want to make AI videos with Google" or "how do I use that Google image AI thing".

0
installs
▌
Vyapar Migration · krishamaze bundle
Authoritative Vyapar schema knowledge. Use when the user mentions Vyapar, .vyb files, .vyp SQLite databases, reading Vyapar data, or any task that requires understanding Vyapar's internal structure. Triggers on: "vyapar", "vyb file", "vyp file", "vyapar backup", "vyapar schema", "vyapar sqlite", "vyapar data". This skill contains only verified facts about Vyapar's internals -- zero assumptions, zero hallucination. Use it regardless of the target system (Odoo, Tally, custom ERP, CSV export, or anything else).

0
installs
▌
Humanization 2026 · krishamaze bundle
Advanced 2026 browser automation humanization — mouse trajectories, keystroke dynamics, scroll behavior, session pacing, and anti-detection evasion. Use this skill whenever writing any Playwright or Camoufox interaction code that must avoid bot detection: typing, clicking, scrolling, navigating, or managing session behavior. DO NOT write bare page.click() or page.fill() without applying humanization from this skill. Covers the full science: Bézier + Fitts's Law mouse curves, Gaussian keystroke timing, scroll jitter, session warmup, and what 2026 ML detectors actually measure.

0
installs
▌
Codex Orchestrator · krishamaze bundle
Authoritative guide for making an LLM fully operate OpenAI Codex CLI as its assistant — replacing the human at the keyboard. Covers all three orchestration tiers: codex exec (simple automation), codex app-server JSON-RPC (full stateful multi-turn control), and command/exec/write (interactive stdin / PTY passthrough for y/n prompts, passwords, OTP). Use this skill whenever the user mentions: "automate Codex", "LLM controls Codex", "outer LLM + Codex", "Codex orchestration", "programmatic Codex", "non-human Codex", or "LLM operate CLI". Do NOT use raw PTY injection into the Codex TUI — that is brittle and confirmed broken by upstream (see references/pty-tui-wall.md).

0
installs
▌
Skill Creator Lean · krishamaze bundle
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. Also trigger when user says "make a skill", "turn this into a skill", "package this", "write a SKILL.md", or asks how to structure agent knowledge — even if they don't say "skill" explicitly. Enforces the map pattern: SKILL.md is an index of pointers, not a knowledge dump. All skills produced must use [REMIND], [READ], [RUN], [SPAWN], [DO] pointer syntax.

0
installs
▌
Private API Reversal · krishamaze
Methodology for reverse-engineering the private/undocumented HTTP API of any web application and building a programmatic client that replays those requests. Use this skill whenever a user wants to automate a web product that has no public API, build an unofficial client for a web service, intercept and replay browser requests programmatically, or asks how to "do what notebooklm-py did" for any other product (Flow, Perplexity, CapCut, etc.). Also trigger for questions about extracting auth tokens from browsers, decoding private API request formats, handling CSRF tokens, or building TypeScript/Python clients around undocumented endpoints. Trigger even for partial phrases like "automate X without an API", "reverse engineer X", or "how does X talk to its backend".

0
installs
▌
Google Flow Prompting · krishamaze
Advanced, non-obvious prompting techniques specific to Google Flow's Nano Banana 2 (image) and Veo 3.1 (video) models. Use this skill whenever a user wants to get better results from Flow, fix specific failure modes (face drift, illustration vs photo, text artifacts, audio mismatch, subject inconsistency), or asks how to prompt Nano Banana, Veo 3.1, or Flow for production-quality output. Trigger even for questions phrased as "why does my Flow generation keep doing X" or "how do I make Nano Banana do Y". Contains only model-specific knowledge that generic prompting guides and LLMs do not have — skip this skill if the user only needs general prompting concepts.

0
installs
▌
Nextjs Approuter 2026 · krishamaze
Next.js App Router patterns for 2026 (Next.js 15+/16+). Use this skill when building the web dashboard frontend — layouts, server components, client components, route handlers, server actions, and data fetching. NEVER use Pages Router, getServerSideProps, or getStaticProps. ALWAYS default to Server Components unless interactivity is needed. Covers the 'use client'/'use server' boundary, fetch caching, and Tailwind v4.

0
installs
▌
Playwright Async 2026 · krishamaze
Playwright async Python API patterns for 2026. Use this skill whenever writing browser automation code in Python — page navigation, element interaction, selectors, persistent context, storage state, and event handling. This is the UNDERLYING API that Camoufox wraps. Know this to understand what Camoufox exposes. ALWAYS use async_api (not sync_api) when integrating with FastAPI or any async Python service.

0
installs
▌
Writing Agent Prompts · krishamaze bundle
Patterns for writing effective prompts for Claude Code (the VS Code / terminal agent). Provides verified patterns for writing effective prompts for Claude Code agents. ALWAYS load this skill when: writing a task for Claude Code, crafting an agent instruction, designing a subagent prompt, debugging Claude drift or incorrect output, reviewing a CLAUDE.md file, structuring a new skill, or any question about how to instruct Claude Code effectively. Trigger phrases: "prompt for claude", "tell claude to", "claude drifted", "claude did wrong", "how to ask claude", "agent task", "CLAUDE.md", "write a skill", "skill description", "claude ignored", "claude did the wrong thing".

0
installs
▌
Observability Engineering · krishamaze bundle
Deep observability engineering knowledge from "Observability Engineering" (Majors, Fong-Jones, Miranda). Read this skill whenever Claude Code is asked to: instrument code for observability, debug production issues, set up or query monitoring/tracing/logging tools, configure SLOs or alerts, design telemetry pipelines, review instrumentation quality, or investigate incidents. Also triggers on keywords: observability, o11y, tracing, spans, cardinality, SLO, SLI, error budget, structured events, OpenTelemetry, OTel, sampling, telemetry pipeline, BubbleUp, core analysis loop, wide events, MELT (metrics/events/logs/traces), Honeycomb, Datadog, Grafana, Prometheus, Loki, Tempo, Jaeger. This skill is the WHAT (book wisdom) that pairs with MCP tools as the HOW (live data access). Always read this before writing instrumentation code, querying observability backends, or advising on monitoring strategy — even if the user doesn't explicitly mention "observability."

0
installs

21 published skills