ClawHub

Agent Architect

v1.0.2

Interactive consultant that helps developers design agent systems. Walks through structured intake questions about surfaces, tools, memory, deployment, and c...

⭐ 0· 60·0 current·0 all-time
by@philitician
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
The name/description (agent architecture consultant) matches the runtime instructions and included reference files. All references and the intake/synthesis flow are coherent with designing agent systems; no unrelated privileges, binaries, or credentials are requested.
✓
Instruction Scope
SKILL.md limits behavior to a structured Q&A intake, reading the provided reference files, and synthesizing recommendations. It does not instruct the agent to read arbitrary system files, environment variables, network endpoints, or transmit data externally. The requirement to 'read relevant files before making claims' refers to the local bundled references.
✓
Install Mechanism
There is no install spec and no code files — the skill is instruction-only, which minimizes risk because nothing is written to disk or executed beyond the agent's normal prompt processing.
✓
Credentials
The skill requests no environment variables, credentials, or config paths. The scope of required access is proportional to its function (reading local reference markdown files).
✓
Persistence & Privilege
The skill does not request 'always: true' or system-level modifications. It is user-invocable and allows autonomous model invocation (platform default), which is appropriate for a consultative skill and not by itself a concern.
Assessment
This instruction-only skill appears coherent and low-risk: it asks the agent to walk a user through intake questions and to ground recommendations in the bundled reference markdown files. Before installing, you may want to (1) quickly inspect the included reference files to ensure they contain only documentation (no embedded secrets or unexpected endpoints), and (2) confirm you trust the skill author/source since the package metadata lacks a homepage. Autonomous invocation is allowed (normal default); if you prefer to limit autonomous runs, control the agent's skill permissions in your platform settings before enabling.

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

latestvk970w574vtbnngmy5xkcq4tw8d84chf2

License

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

Comments