ClawHub

Agentic Coding

v1.0.0

Ship production code with AI agents through acceptance contracts, micro diffs, red green loops, and deterministic handoff checkpoints.

⭐ 7· 8.5k·113 current·118 all-time
byIván@ivangdavila
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name/description (ship production code with contract-driven loops) align with required binary (git) and the instruction set (create diffs, run tests, produce handoffs). Required resources and declared scope are proportional to code-delivery workflows, including Xcode-specific guidance for iOS/macOS work.
ℹ
Instruction Scope
SKILL.md and companion docs confine reads/writes to ~/agentic-coding/ and the current repository context and explicitly state no undeclared network requests; this is coherent. Note: the skill is instruction-only — the agent executing these instructions could still run repo-local commands (tests, builds) which may have side effects; the docs rely on agent compliance for 'no network' guarantees.
✓
Install Mechanism
No install spec or external downloads — instruction-only skill. This has minimal on-disk footprint (creates files under the user's home path when the user opts in) and raises low installation risk.
✓
Credentials
No environment variables, credentials, or external config paths are requested. The single required binary (git) is appropriate for producing diffs and handoffs; no unrelated secrets or services are required.
ℹ
Persistence & Privilege
The skill persists its memory and artifacts under ~/agentic-coding/ if the user opts in. always:false (default) and no platform/global config changes are requested. Users should be aware these local files will be created and may contain contract text and validation evidence.
Assessment
This instruction-only skill appears coherent and safe for its stated purpose. Before installing: (1) ensure you are comfortable with the skill creating and reading files under ~/agentic-coding/ (do not put secrets there); (2) confirm git is available and that you are happy for the agent to run repo-local commands and tests (these can have side effects like builds or simulator launches); (3) understand that the SKILL.md claims no undeclared network requests, but that is an operational constraint the agent must follow — monitor initial runs and review the created memory/hand-off files to verify behavior. If you need stricter guarantees, run the skill in a sandboxed repo or review its produced files before merging changes.

Like a lobster shell, security has layers — review code before you run it.

latestvk9740y2cktdz3zk2pha9frvawh81zd9d

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Runtime requirements

AI Clawdis
OSLinux · macOS · Windows
Binsgit

Comments