Metric Definition Catalog
v1.0.0把散落指标统一整理成口径、公式、归属、例外情况与常见误用。；use for metrics, catalog, analytics workflows；do not use for 编造指标来源, 替代 BI 平台配置.
⭐ 0· 92·0 current·0 all-time
byvx：17605205782@52yuanchangxing
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description, SKILL.md, resources, and the included scripts all align: the bundle is designed to take metric-related inputs (files or text) and produce structured Markdown reports or audits. Requested binary (python3) and declared file resources are proportionate to the stated functionality.
Instruction Scope
SKILL.md instructs running scripts/run.py to produce reports and to prefer read-only/dry-run outputs. The script legitimately reads files, directories, and CSVs and can perform directory-level scans and pattern checks. This is coherent for analysis/audit use, but it means the skill can read arbitrary user-supplied file paths (including code, configs, CSVs)—so users should not point it at system or secret-containing directories unless they intend that.
Install Mechanism
No install spec or external downloads; this is an instruction-plus-local-script skill that only requires an existing python3 on PATH. No network fetches or archive extractions are performed by the skill package itself.
Credentials
No environment variables, secrets, or external credentials are requested. The script includes secret-detection regexes (to flag possible secrets) but does not contain or require credentials itself.
Persistence & Privilege
Skill is user-invocable, not forced-always, and does not request persistent agent-wide privileges or attempt to modify other skills. It can be invoked autonomously per platform defaults, which is expected behavior and not in itself a concern.
Assessment
This skill appears to be what it says: a local tool to structure metric definitions and audit related files. Before installing or running it, review scripts/run.py yourself and run it in a controlled workspace. Do not point the script at system roots or directories containing secrets (e.g., /etc, your home dir, cloud credential files) unless you intentionally want those files scanned. Prefer --dry-run and output to a sandbox file, and inspect generated output for any redacted or sensitive content before sharing. If you need stronger isolation, run the script in a container or VM. If you want confirmation, request the author/publisher identity or run the included tests/smoke-test locally first.Like a lobster shell, security has layers — review code before you run it.
latestvk971yz7j4k899j7nfkwbc09p4s835gyw
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
📐 Clawdis
OSmacOS · Linux · Windows
Binspython3