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 agentmods add commands/druide67/asiai/detectgit clone --depth 1 https://github.com/druide67/asiaiWhat 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 | $0.00025 | $0.00544 |
| Opus 5 | $0.00013 | $0.00272 |
| Sonnet 5 | $0.00005 | $0.00109 |
| Haiku 4.5 | $0.00003 | $0.00054 |
Grade A, and why
detect 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 3d 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.
What it actually says
asiai detect
Auto-detect running inference engines using a 3-layer cascade.
Usage
asiai detect # Auto-detect (3-layer cascade)
asiai detect --url http://host:port # Scan specific URL(s) only
Output
Detected engines:
● ollama 0.17.4
URL: http://localhost:11434
● lmstudio 0.4.5
URL: http://localhost:1234
Running: 1 model(s)
- qwen3.5-35b-a3b MLX
● omlx 0.9.2
URL: http://localhost:8800
How it works: 3-layer detection
asiai uses a cascade of three detection layers, from fastest to most thorough:
Layer 1: Config (fastest, ~100ms)
Reads ~/.config/asiai/engines.json — engines discovered in previous runs. This catches engines on non-standard ports (e.g., oMLX on 8800) without rescanning.
Layer 2: Port scan (~200ms)
Scans default ports plus an extended range:
| Port | Engine |
|---|---|
| 11434 | Ollama |
| 1234 | LM Studio |
| 8080 | mlx-lm or llama.cpp |
| 8000-8009 | oMLX or vllm-mlx |
| 52415 | Exo |
Layer 3: Process detection (fallback)
Uses ps and lsof to find engine processes listening on any port. Catches engines running on completely unexpected ports.
Auto-persist
Any engine discovered in Layer 2 or 3 is automatically saved to the config file (Layer 1) for faster detection next time. Auto-discovered entries are pruned after 7 days of inactivity.
When multiple engines share a port (e.g., mlx-lm and llama.cpp on 8080), asiai uses API endpoint probing to identify the correct engine.
Explicit URLs
When using --url, only the specified URLs are scanned. No config is read or written — useful for one-off checks.
asiai detect --url http://192.0.2.10:11434,http://localhost:8800
See also
- config — Manage persistent engine configuration
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.
- 3d ago First seen · 74 lines · 25 tokens per session scan A 1dcfafb255d7
detect is a command published in the GitHub repository druide67/asiai (11 stars, last pushed 3d ago), licensed Apache-2.0. It adds 25 tokens to every session and 544 once invoked, about $0.0001 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 2026-08-30.
Other commands, from other repositories
perfup
Autonomous performance optimization: research, PoC, benchmark, implement, review, PR.
memo
Search, save, inspect, and maintain memo persistent memory.
verify
Spawn a fresh-context verifier subagent to check completed work against its specification before trusting it.
setup
Install or verify the ADE-Bench harness at /.ade-bench (clones repo, installs CLI, downloads bundled DuckDB databases).
/opsx-apply
Implement tasks from an OpenSpec change (Experimental).
/opsx-sync
Sync delta specs from a change to main specs.