Install
$ agentstack add skill-ewanjohndennis-engineering-standard-skills-engineering-standard-skills ✓ 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.
About
Engineering Standards
> Inspired by Google's Engineering Practices documentation (CC-BY 3.0) and general production engineering norms. Not affiliated with or endorsed by Google. See README for attribution details.
This skill bundles several layers of engineering guidance. Don't apply everything blindly — read in only the reference files relevant to the current task, then apply them silently. Never narrate which files you're reading or that you're "loading a skill."
How to route
Step 1 — Is this worth calibrating?
If the task is non-trivial (a real feature, a refactor, a full review, a design question) — not a ten-line fix or a quick question — read references/calibration.md first and ask its one calibration question before producing anything. Skip this for small tasks or when the user has already given context (e.g. "this is a quick prototype," "this is going to prod").
Step 2 — Generation or review?
- If the user wants code written, built, refactored, or debugged: read
references/code-generation.mdand apply it. - If the user wants existing code, a diff, or a PR reviewed or critiqued: read
references/code-review.mdand apply it. - If the task involves both (e.g. "write this, then review it"), read both.
Step 3 — Which language?
Detect the language from the code, the file extension, or the user's request. Read the matching reference file and apply it on top of the generation/review standards:
- Python →
references/python-standards.md - Java →
references/java-standards.md - C →
references/c-standards.md
If the language isn't one of these three, just apply the generic generation/review standards — they're language-agnostic by design and already defer to whatever conventions the codebase uses.
If a single task spans multiple languages (e.g. a Python backend with embedded SQL, or a C extension called from Python), read the reference files for each language actually present.
Priority order across all layers
When guidance from different reference files conflicts, resolve it in this order:
- Correctness
- Security
- Existing codebase conventions
- Maintainability
- Readability
- Performance
- Style preferences
A language-specific idiom never overrides a higher-priority concern from the generic standards. Existing codebase conventions outrank every default in every reference file here.
Output behavior
Apply all loaded guidance silently — no preamble like "Based on the engineering-standards skill, I will..." Just produce the code, the review, or the calibration question, as appropriate. If the user pushes back on feedback, handle it the way references/code-review.md describes.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Ewanjohndennis
- Source: Ewanjohndennis/engineering-standard-skills
- 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.