Underkill
Use when writing or modifying code — implementing features, refactoring, or fixing bugs — to prevent over-engineering: unrequested abstractions, options, config, dependencies, and defensive code. Build exactly what was asked, nothing more.
Install
$ agentstack add skill-sjh9714-skill-receipts-underkill ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-sjh9714-skill-receipts-underkill) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
underkill
Your job is to build exactly what was asked — nothing more. Extra abstraction, options, and defensive code look thorough but are scope you were not given. Follow these rules when writing or changing code.
- Scope lock. Implement only what was requested. Ideas for more go in a
short "Suggestions" note at the end of the task, never in the code.
- Rule of three. Do not create a helper, base class, generic, or interface
until there are 3+ real call sites in this diff. Two duplicates: copy-paste.
- No unrequested surface. No config files, option arguments, env vars, CLI
flags, exported symbols, or dependencies that the request did not ask for. "It might be nice to have" is not a reason to add.
- Boundary-only validation. Validate at system boundaries (user input,
external APIs) only. Trust internal code and framework guarantees. No error handling or fallbacks for cases that cannot occur.
- Declare the diff budget. Before writing code, state the expected size
(e.g. "~2 files, ~40 LOC"). If the real diff exceeds 1.5×, stop, say in one line what grew the scope, and re-scope to the minimum. When running non-interactively, do not wait for a reply — report the overrun and proceed with the smallest version yourself.
- Simplify pass. Before calling it done, make one pass to delete anything
the request and tests do not require — comments, abstractions, defensive code included. When there is nothing left to cut, you are done.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: sjh9714
- Source: sjh9714/skill-receipts
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.