Verify Config
ASecurityConfig doctor for clawback — loads the merged config (DEFAULTS < global < ./CLAWBACK.md < CLI overrides) exactly as the proxy's loadConfig does, prints a secret-free summary of the resolved values plus the merge-order sources, and exits non-zero if the canonical CLAWBACK.md did NOT take effect (host not 0.0.0.0, tls off, or adminToken missing) so a bad parse fails loudly instead of silently falling back to loopback defaults. Read-only; never prints the adminToken value, only its length. Use t...
- 5 stars
- 0 votes
- 0 copies
- 2 views
- Added August 31, 2026
Works with
Security analysis
100/100Pro scans all 2 files and shows the line behind each finding
npx -y skills add zapgun-ai/clawback --skill verify-config --agent claude-codeAre you the author of Verify Config?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/zapgun-ai-verify-config)---
name: verify-config
description: Config doctor for clawback — loads the merged config (DEFAULTS < global < ./CLAWBACK.md < CLI overrides) exactly as the proxy's loadConfig does, prints a secret-free summary of the resolved values plus the merge-order sources, and exits non-zero if the canonical CLAWBACK.md did NOT take effect (host not 0.0.0.0, tls off, or adminToken missing) so a bad parse fails loudly instead of silently falling back to loopback defaults. Read-only; never prints the adminToken value, only its length. Use to confirm a CLAWBACK.md edit actually parsed and merged before starting the proxy, or to debug which config layer won.
---
# clawback config doctor
Run `.skills/verify-config/scripts/verify_config_parse.mjs` from the project
root. It loads the merged config the same way the proxy would and prints what
actually resolved — no secrets — then fails loudly if the canonical
`CLAWBACK.md` did not take effect.
```bash
node .skills/verify-config/scripts/verify_config_parse.mjs # resolve from ./ (cwd)
node .skills/verify-config/scripts/verify_config_parse.mjs /path/to/project
```
The optional positional argument is the directory to resolve config from
(defaults to the current working directory).
## What it prints
- **sources (merge order)** — each contributing tier and file path, lowest
precedence first, exactly the layering in CLAUDE.md's *Config Merging
Strategy* (`DEFAULTS < global < local < CLI`).
- **resolved** — host, port, tls, cert/key paths, selfSign, passthrough,
keepAliveEnabled, injectExtendedCacheTtl, adminPathPrefix, and whether an
adminToken is present (length only — the value is never shown).
- **warnings** — any non-fatal notes `loadConfig` surfaced.
## Exit status
Exits **non-zero** with a `FAIL:` list when any of these are off — the
signature of a `CLAWBACK.md` that did not parse or merge:
- `host` is not `0.0.0.0` (the canonical LAN bind did not stick),
- `tls` is not `true`,
- `adminToken` is missing (a non-loopback bind would have thrown anyway).
Otherwise it prints `OK: canonical CLAWBACK.md parsed and merged`. Pairs with
the [write-config](../write-config/SKILL.md) skill, which generates that file.
Files in this skill
- SKILL.md
- scripts/verify_config_parse.mjs
Attribution
Comments
Loading comments…
Newsletter
Get the weekly list of safe new skills, and the ones we pulled.