Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/howdeploy/deploychan_mcpnpx agentmods add skills/howdeploy/deploychan_mcp/zaebalWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/howdeploy/deploychan_mcp/zaebal)<a href="https://agentmods.dev/skills/howdeploy/deploychan_mcp/zaebal"><img src="https://agentmods.dev/badge/skills/howdeploy/deploychan_mcp/zaebal/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/howdeploy/deploychan_mcp/zaebal"><img src="https://agentmods.dev/badge/skills/howdeploy/deploychan_mcp/zaebal.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Agent Snooping · line 100 Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.Fix: Remove all code or instructions that access agent configuration directories (.claude/, .codex/, .gemini/). If configuration values are needed, pass them explicitly as parameters or environment variabl
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00085 | $0.02655 |
| Opus 5.5 | $0.00034 | $0.01062 |
| Sonnet 5.5 | $0.00017 | $0.00531 |
| Haiku 4.5 | $0.00009 | $0.00265 |
Grade A, and why
Z.A.E.B.A.L.: profanity-triggered self-audit for agents scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directorieslowAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
| Claude Code | Existing Linux adapter; `~/.claude/settings.json` | `claude -p`, read-only tools | Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
How it starts
The opening of the file, as written. The whole thing — 202 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Z.A.E.B.A.L. — profanity-triggered self-audit
Zaebal? Audit. Errors. Break. Analyze. Leave no assumption.
When a coding agent gets stuck, it repeats the same action with small variations — because
one underlying belief about the task or the codebase is wrong, and the agent treats that
belief as fact. Another self-check just reproduces the mistake. Z.A.E.B.A.L. plugs a
feedback loop into the user-message boundary: profanity and direct complaints become an
audit signal, repeated signals escalate, and at the top level an external auditor
(the same or a cross-vendor CLI) reads the transcript and repository evidence from outside.
Public: howdeploy/Z.A.E.B.A.L, MIT, Python standard library only.
The upstream installation router and Windows/Linux change are the source for the platform instructions below. Fetch the skill and runtime from the same repository revision; downloading just the skill folder does not install a working hook.
Why it works
The named failure patterns (collected from real session postmortems):
- The wrong belief is invisible to the agent. It doesn't loop from inattention — it has sincerely stopped understanding the problem. The protocol's goal is not "find the mismatch" but find and disprove the wrong belief.
- "Written ≠ took effect". A classic: the agent wrote a config/hook/instruction file and assumes it works because the file exists — while the harness reads it from a different path. Verify the act of consumption, not the act of writing.
- Sycophancy vs hallucinated correctness. Agreeing under pressure and abandoning a working solution — or defending broken code with invented facts. The cure for both: execution over intuition. Defending code with verbal arguments is forbidden — only a micro-test, a run, logs.
- First plausible hypothesis. A lone agent fixates on the first version. That's why auditors get raw artifacts, not the agent's interpretation — and why there are two of them: independent versions disprove each other's dead ends.
- Missing session history. Read the original request through the trigger, locate the first divergence, and correlate it with diffs, commits and actual test/error output. A transcript excerpt is orientation; read the full source when its path is available. Mark causal conclusions unverified when missing history could change them.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed · +86 lines · +14 tokens per session ab0f6fdfd371
- 1mo ago First seen · 116 lines · 71 tokens per session scan A ecd855de3ae7
Z.A.E.B.A.L.: profanity-triggered self-audit for agents is a skill published in the GitHub repository howdeploy/deploychan_mcp (18 stars, last pushed 3d ago), licensed MIT. It adds 85 tokens to every session and 2,655 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
company-handbook
The cross-director coordination protocol for the 1ai business kingdom. Defines how Content, Video, Image, Marketing, and Sales directors work together. Has memory via 1ai-hub brain, follows 1ai-rules, and coordinates with Vilona (GM). Use when coordinating multiple directors, resolving conflicts between teams, or when…
linux-networking
Linux networking expert for iptables, nftables, routing, DNS, and network troubleshooting.
sysadmin
System administration expert for Linux, macOS, Windows, services, and monitoring.
triage-issues
Triage GitHub issues in the googleapis/mcp-toolbox repo: propose the correct labels (type / priority / product / status), check for duplicates, verify a bug has enough info to act on, and draft a triage comment. Use whenever a maintainer asks you to triage, label, categorize, prioritize, or "look at" an issue (or a…
adk-debug
Diagnoses misbehaving ADK agents by inspecting sessions, events, tool calls, and the exact request that reached the model. Covers the adk run CLI and the adk web dev server with its session, trace, and debug HTTP endpoints. Use when an agent returns the wrong answer, ignores a tool or swallows a tool error, hangs…
cvm-ai-doctor
A health-diagnosis workflow for servers, computers, virtual machines, and containers on Linux, macOS, or Windows.