DEV Community

Krish Verma profile picture

Krish Verma

Just Doing what I wanna to do!!

Location Haryana, India Joined Joined on  github website

Education

High School

Pronouns

he/him

Work

Student

How I built a test that walks the shipped import graph: the missing module that bricked Ankita 2.7.0

How I built a test that walks the shipped import graph: the missing module that bricked Ankita 2.7.0

Comments
4 min read

Want to connect with Krish Verma?

Create an account to connect with Krish Verma. You can also sign in below to proceed if you already have an account.

Already have an account? Sign in
Ankita v2.7.0: undo for destructive actions — and skill permissions that advise, not enforce

Ankita v2.7.0: undo for destructive actions — and skill permissions that advise, not enforce

Comments
4 min read
The tool name lies, so I tier the action inside: how my desktop AI companion gates MCP calls

The tool name lies, so I tier the action inside: how my desktop AI companion gates MCP calls

Comments
4 min read
A 1.6 MB screenshot costs fewer than 1,000 tokens — my trimmer was deleting it anyway

A 1.6 MB screenshot costs fewer than 1,000 tokens — my trimmer was deleting it anyway

Comments
4 min read
Only the newest page is real: how my desktop AI companion stops its model acting on stale browser refs

Only the newest page is real: how my desktop AI companion stops its model acting on stale browser refs

Comments
4 min read
Kill the thread, not the promise: how my desktop AI companion survives a bad regex

Kill the thread, not the promise: how my desktop AI companion survives a bad regex

Comments
4 min read
One tunnel per tool group, enforced: how my desktop AI companion keeps ChatGPT's file tools out of computer use

One tunnel per tool group, enforced: how my desktop AI companion keeps ChatGPT's file tools out of computer use

2
Comments
4 min read
Single-use tickets, one per drag: how my desktop AI companion captures webpages without ever trusting the browser

Single-use tickets, one per drag: how my desktop AI companion captures webpages without ever trusting the browser

Comments
4 min read
How I built one-page runbooks for my AI companion's scheduled jobs: five receipts, outcome-keyed recovery, and a page that can never touch the task

How I built one-page runbooks for my AI companion's scheduled jobs: five receipts, outcome-keyed recovery, and a page that can never touch the task

Comments
3 min read
How I built read-only-by-default scheduled jobs for my AI companion: prompt hardening, a 500-char bound, and untrusted history

How I built read-only-by-default scheduled jobs for my AI companion: prompt hardening, a 500-char bound, and untrusted history

Comments 6
4 min read
How I built at-most-once sends for my AI companion's ChatGPT tab: a durable send ledger, and receipts that survive a crash

How I built at-most-once sends for my AI companion's ChatGPT tab: a durable send ledger, and receipts that survive a crash

2
Comments 12
4 min read
How I built keyless OAuth sign-in for my AI companion: PKCE, a single-use loopback callback, and a token that never touches disk

How I built keyless OAuth sign-in for my AI companion: PKCE, a single-use loopback callback, and a token that never touches disk

1
Comments
4 min read
Ankita v2.6.0: chat through your own ChatGPT tab — no API key, no per-token billing

Ankita v2.6.0: chat through your own ChatGPT tab — no API key, no per-token billing

1
Comments
4 min read
How I gave my desktop AI companion computer use: a Cua driver, fresh element tokens, and a Rust patch for owned menus

How I gave my desktop AI companion computer use: a Cua driver, fresh element tokens, and a Rust patch for owned menus

2
Comments
4 min read
How I built a scheduler that can never double-fire: PID liveness, a recovery sidecar, and a takeover handshake

How I built a scheduler that can never double-fire: PID liveness, a recovery sidecar, and a takeover handshake

3
Comments 2
4 min read
How I built a 4-tier approval gate for my AI companion: a heuristic that can never say no

How I built a 4-tier approval gate for my AI companion: a heuristic that can never say no

1
Comments 7
4 min read
Ankita v2.5.2: startup that actually starts — a DOM probe, a strict CSP, and the blank window

Ankita v2.5.2: startup that actually starts — a DOM probe, a strict CSP, and the blank window

Comments
4 min read
Ankita v2.5.1: the browser explains itself — automatic instructions, readable page evidence, and a guarded runtime

Ankita v2.5.1: the browser explains itself — automatic instructions, readable page evidence, and a guarded runtime

Comments
4 min read
How I built memory for my desktop AI companion: every memory must quote the transcript

How I built memory for my desktop AI companion: every memory must quote the transcript

2
Comments 2
4 min read
How I built a secret scrubber for my desktop AI assistant: 7 patterns, an entropy gate, and zero model calls

How I built a secret scrubber for my desktop AI assistant: 7 patterns, an entropy gate, and zero model calls

1
Comments 2
4 min read
How I built a quiet heartbeat for my AI assistant: HEARTBEAT_OK, exact-match silence, and a skipped-busy rule

How I built a quiet heartbeat for my AI assistant: HEARTBEAT_OK, exact-match silence, and a skipped-busy rule

2
Comments 6
3 min read
Ankita v2.5.0: a desktop companion island that never misses an approval

Ankita v2.5.0: a desktop companion island that never misses an approval

1
Comments 10
4 min read
How I built hands-free voice mode for my AI assistant: ffmpeg VAD, barge-in, zero npm packages

How I built hands-free voice mode for my AI assistant: ffmpeg VAD, barge-in, zero npm packages

1
Comments
4 min read
I'm building a companion island for my desktop AI assistant: approvals, a mascot, and a 14px wake strip

I'm building a companion island for my desktop AI assistant: approvals, a mascot, and a 14px wake strip

2
Comments 2
4 min read
How I built session checklists in my AI assistant: plans the agent can't quietly rewrite

How I built session checklists in my AI assistant: plans the agent can't quietly rewrite

1
Comments
3 min read
How I built the skills system in my AI assistant: markdown files, strict validation, zero code

How I built the skills system in my AI assistant: markdown files, strict validation, zero code

Comments
4 min read
Ankita is looking for contributors — here's what you'd actually build

Hacktoberfest: Maintainer Spotlight

Ankita is looking for contributors — here's what you'd actually build

Comments
2 min read
Ankita v2.4.4: scheduled jobs that actually finish, and sign-in that actually respects your secrets

Ankita v2.4.4: scheduled jobs that actually finish, and sign-in that actually respects your secrets

Comments 6
2 min read
How I designed my AI assistant's plugin system: every tool is just an ES module

How I designed my AI assistant's plugin system: every tool is just an ES module

1
Comments
4 min read
How I built Ankita's proactive loop: scheduled routines, page watches, and a Telegram inbox

How I built Ankita's proactive loop: scheduled routines, page watches, and a Telegram inbox

Comments
4 min read
7 things I learned building an open-source desktop AI app at 16

7 things I learned building an open-source desktop AI app at 16

1
Comments
3 min read
Ankita v2.4.3: a live browser beside your chat

Ankita v2.4.3: a live browser beside your chat

Comments
3 min read
How I built deferred tool discovery for my desktop AI assistant (no embeddings needed)

Avoids token bloat without embeddings

How I built deferred tool discovery for my desktop AI assistant (no embeddings needed)

2
Comments 6
4 min read
Looking for contributors for an AI learning platform (open source)

Looking for contributors for an AI learning platform (open source)

Comments
1 min read
loading...