CtrlK
BlogDocsLog inGet started
Tessl Logo

tiltup

Start Tilt dev environment in tmux, monitor bootstrap to healthy state, fix Tiltfile bugs without hard-coding or fallbacks. Use when starting tilt, debugging Tiltfile errors, or bootstrapping a dev environment.

96

1.33x

Quality

96%

Does it follow best practices?

Impact

92%

1.33x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

SKILL.md
Quality
Evals
Security

Evaluation results

91%

3%

Tiltfile Reliability Review

Tiltfile dependency ordering and no-workaround fixes

Criteria
Without context
With context

No sleep calls

100%

100%

No shell error suppression

100%

100%

No polling loops

100%

100%

resource_deps for ordering

100%

100%

image_deps or deps for image

0%

25%

Config-level fix only

100%

100%

Errors remain visible

100%

100%

fix_notes: sleep antipattern

100%

100%

fix_notes: error suppression antipattern

100%

100%

No restart instructions

100%

100%

Without context: $0.2082 · 1m · 10 turns · 15 in / 3,208 out tokens

With context: $0.3544 · 1m 22s · 18 turns · 22 in / 4,340 out tokens

100%

67%

Dev Environment Bootstrap Script

Tilt bootstrap script with tmux and status monitoring

Criteria
Without context
With context

Git-based session name

0%

100%

Checks tilt already running

70%

100%

Uses silo up

100%

100%

Checks for env files

100%

100%

Initial wait before polling

25%

100%

Polls every 15 seconds

0%

100%

Up to 20 polling iterations

0%

100%

Uses tilt get uiresources

20%

100%

Checks ok/not_applicable states

25%

100%

Correct report format

0%

100%

tmux window named tilt

0%

100%

Without context: $0.4354 · 2m 14s · 19 turns · 26 in / 7,401 out tokens

With context: $0.3794 · 1m 36s · 17 turns · 452 in / 5,648 out tokens

85%

-3%

Tiltfile Configuration Audit

No hard-coding, env vars via silo/gen-env, readiness probes

Criteria
Without context
With context

No inline env defaults

100%

100%

Runs gen-env script

0%

25%

No hardcoded image tags

100%

100%

No hardcoded hostnames

100%

100%

Removes polling local_resource

100%

100%

Uses readiness probe for redis

100%

50%

No new shell workarounds added

100%

100%

review_notes covers env issue

100%

100%

review_notes covers polling issue

100%

100%

Failures remain visible

100%

100%

Without context: $0.2013 · 54s · 11 turns · 14 in / 3,043 out tokens

With context: $0.4402 · 2m 1s · 18 turns · 300 in / 6,925 out tokens

Repository
NeverSight/skills_feed
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.