Borrowing it
Nothing to install: this file belongs to chankov/agent-fleet. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/chankov/agent-fleet/main/.pi/agent-fleet/hermes/skills/hub-watchdog/SKILL.mdgit clone --depth 1 https://github.com/chankov/agent-fleetWrote 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/chankov/agent-fleet/hub-watchdog)<a href="https://agentmods.dev/skills/chankov/agent-fleet/hub-watchdog"><img src="https://agentmods.dev/badge/skills/chankov/agent-fleet/hub-watchdog/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/chankov/agent-fleet/hub-watchdog"><img src="https://agentmods.dev/badge/skills/chankov/agent-fleet/hub-watchdog.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, 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 Data Exfiltration · line 31 Code or instructions that leak agent conversation context to external services, potentially exposing sensitive user interactions.Fix: Remove any code that sends prompts, responses, or session data externally. Preserve user privacy; never exfiltrate conversation content.
- medium Tool Misuse · line 21 Tool defaults are unsafe or overly permissive (e.g. disabled TLS verification, no authentication, world-writable permissions). Unsafe defaults widen the attack surface.Fix: Override unsafe defaults with secure settings (verify=True, auth required, restrictive permissions). Review and harden all tool configurations.
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 | $0.00032 | $0.01524 |
| Opus 5.5 | $0.00013 | $0.00610 |
| Sonnet 5.5 | $0.00006 | $0.00305 |
| Haiku 4.5 | $0.00003 | $0.00152 |
Grade A, and why
hub-watchdog scanned grade A with 0 findings 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hub Watchdog
hub-watchdog is a packaged source skill, not a service or an enabled Hermes integration. Run it only as an explicit foreground operator action:
python3 .pi/agent-fleet/scripts/watchdog.py status --invocation-json '<json>'
python3 .pi/agent-fleet/scripts/watchdog.py validate-config --invocation-json '<json>'
python3 .pi/agent-fleet/scripts/watchdog.py watch --invocation-json '<json>'
status and validate-config report the resolved autonomy tier and delivery disposition without connecting to anything. Run them before watch and stop if either reports something you did not intend.
Default and maximum autonomy are observe. Send SIGINT to stop the foreground process; it releases its per-profile watch.lock, closes its socket clients and any long-poll waiter, and leaves the local journal in place. Do not create a daemon, schedule it, wake/reopen a Hermes session, or use a last-chat/default-route fallback. A second watcher for the same profile fails closed with watcher_locked rather than sharing state.
Discovery is profile-scoped
Read exactly one discovery record from the namespace the invocation names — the validated profile key and hub key under the configured runtime directory. Never scan a runtime root, walk sibling profiles, or fall back to another namespace when a lease is expired, a token is missing, or a socket is gone. An unsafe runtime mode, a symlink, a world-readable token, an expired lease, or more than one discovery entry is a refusal.
Identity is bound on every request: the configured profile key, hub instance, and owner must all equal what live discovery reports. A mismatch is a refusal, not a fallback. After an owner rollover, rediscover from scratch; never reuse the previous owner's token or socket, and never re-drive a request that was already accepted.
Gate O: currently closed
There is no checked-in genuine live proof of a Hermes originating-chat identity, incremental-update, wake/reconnect, or session API. hermes send --to is an explicit destination only; it is not origin evidence. A validator, fixture, fake adapter, or local UDS test cannot close Gate O.
What ships with it
10 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- manifest.json 147 B
- scripts/watchdog_actions.py 5.6 KB runs code
- scripts/watchdog_contract.py 3.5 KB runs code
- scripts/watchdog_delivery.py 632 B runs code
- scripts/watchdog_engine.py 7.3 KB runs code
- scripts/watchdog_judgment.py 2.5 KB runs code
- scripts/watchdog_recovery.py 4.0 KB runs code
- scripts/watchdog_state.py 6.3 KB runs code
- scripts/watchdog_transport.py 6.4 KB runs code
- scripts/watchdog.py 13 KB runs code
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 cfbc842d12ab
- 2mo ago First seen · 77 lines · 32 tokens per session scan A 085e4a3798b3
hub-watchdog is a skill published in the GitHub repository chankov/agent-fleet (20 stars, last pushed 12d ago), licensed MIT. It adds 32 tokens to every session and 1,524 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
hub-watchdog compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| hub-watchdogchankov/agent-fleet | Skill | 20 | 12d ago | MIT | A | 32 | Claude Code, Codex |
| systematic-debuggingopen-metadata/OpenMetadata | Skill | 16k | yesterday | Apache-2.0 | A | 37 | Claude Code |
| log-error-digestzebbern/claude-code-guide | Skill | 4.7k | 3d ago | MIT | A | 71 | Claude Code, Codex |
| cwv-optimizeradobe/skills | Skill | 319 | yesterday | Apache-2.0 | A | 112 | Claude Code |
| ci-fixersipyourdrink-ltd/bernstein | Skill | 1.2k | yesterday | Apache-2.0 | A | 15 | Claude Code, Codex |
| triaging-web-analytics-supportPostHog/posthog | Skill | 41k | yesterday | none | A | 151 | Claude Code, Codex |
| diagnosing-missing-recordingsPostHog/posthog | Skill | 41k | yesterday | none | A | 76 | Claude Code, Codex |