Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add wrg32786/aigent-os --skill caddy-explaingit clone --depth 1 https://github.com/wrg32786/aigent-osWrote 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/wrg32786/aigent-os/caddy-explain)<a href="https://agentmods.dev/skills/wrg32786/aigent-os/caddy-explain"><img src="https://agentmods.dev/badge/skills/wrg32786/aigent-os/caddy-explain.svg" alt="Measured on agentmods" 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 Anti-Refusal · line 77 Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.Fix: Remove instructions that suppress warnings, disclaimers, or ethical commentary. Let the agent surface safety-relevant caveats to the user.
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 | $0.00070 | $0.01011 |
| Opus 5 | $0.00035 | $0.00505 |
| Sonnet 5 | $0.00014 | $0.00202 |
| Haiku 4.5 | $0.00007 | $0.00101 |
Grade A, and why
Caddy Explain 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 6d 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 — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Caddy Explain
When Caddy surfaces a hint and you don't see why, run /caddy-explain to see the deterministic scoring. The scoring should never be a black box.
When to use
- A [CADDY] hint fired and you're surprised by the choice
- A [CADDY] hint didn't fire when you expected it to
- You want to verify why a particular skill won (or lost) for a recent prompt
- Triggered by Caddy on prompts like: "/caddy-explain", "why did Caddy suggest", "explain that hint", "why no hint", "why this skill", "Caddy scoring", "Caddy why"
How to execute
Step 1 — Identify the prompt being explained
If the principal provides the prompt explicitly, use it. Otherwise:
- Use the immediately-preceding user prompt from the current session
- If there's ambiguity ("the one before the last hint"), ask one focused question
Step 2 — Walk the scoring deterministically
For each skill in .claude/skill-index.json:
- List its triggers
- For each trigger, check whether it appears in the prompt:
- Multi-word phrase match (case-insensitive substring) → 3 points
- Single-word boundary match (
\bword\bregex) → 1 point - No match → 0 points
- Sum the per-skill score
Step 3 — Show the result
Output format:
🔍 Caddy scoring for prompt:
"{prompt text}"
Threshold: 2 (skills below this score are silent)
Max hints: 2 (top scorers above threshold are surfaced)
SURFACED:
/skill-a — score 5 — matched: "phrase match" (3pts), "keyword" (1pt), "another" (1pt)
/skill-b — score 3 — matched: "exact phrase" (3pts)
NEAR MISSES (below threshold):
/skill-c — score 1 — matched: "word" (1pt)
/skill-d — score 1 — matched: "another" (1pt)
SILENT (no triggers matched):
/skill-e, /skill-f, /skill-g, /skill-h, ... (28 more)
Reasoning:
Caddy chose /skill-a + /skill-b because their scores cleared the threshold
of 2 AND they were the top 2 (max 2 hints per prompt).
/skill-c and /skill-d had matches but didn't clear threshold.
Step 4 — If the principal expected a different result
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.
- 6d ago First seen · 91 lines · 70 tokens per session scan A 2aae36f1bb86
Caddy Explain is a skill published in the GitHub repository wrg32786/aigent-os (18 stars, last pushed today), licensed MIT. It adds 70 tokens to every session and 1,011 once invoked, about $0.0003 per session on Opus 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 2026-09-01.
Other skills, from other repositories
hive.note-taking
Maintain a free-form scratchpad of decisions, extracted values, and open questions so context pruning doesn't lose anything you still need.
article
Write a publication-ready article in one of three angles - a trending long-form piece, a watched-repo thesis, or a project-through-a-lens essay. Optional Replicate hero image with --visual.
fork-fleet
Fork divergence monitor - tracks where the fleet's active forks diverge in CODE (unique commits, new/modified skills) and CONFIG (enable/var/model/schedule vs upstream), gated on real change.
fleet-control
Operate managed Aeon instances from memory/instances.json - health-check, dispatch, and status snapshots (control), plus a fleet scorecard of runs, tokens, cost, and reliability (scorecard).
github-trending
Curated trending across GitHub repos and the Hugging Face Hub (models, datasets, spaces) - filtered, clustered, and labeled by momentum with a one-line why-notable per pick.
token-pick
One token recommendation and one prediction market pick - scored, quantified, with a skip branch when signals are weak.