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.
git clone --depth 1 https://github.com/alinaqi/maggyWrote 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/commands/alinaqi/maggy/model-config)<a href="https://agentmods.dev/commands/alinaqi/maggy/model-config"><img src="https://agentmods.dev/badge/commands/alinaqi/maggy/model-config.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00000 | $0.00661 |
| Opus 5 | $0.00000 | $0.00331 |
| Sonnet 5 | $0.00000 | $0.00132 |
| Haiku 4.5 | $0.00000 | $0.00066 |
Grade B, and why
model-config scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
MR="$(cat ~/.claude/.bootstrap-dir)/scripts/model_routing.py" How it starts
The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/model-config — Choose the model you follow
View or change the primary model that handles your coding work across
srooter (the gateway), the route-task hooks, and Maggy. One setting, read
everywhere. Stored in ~/.claude/model-config.json.
Usage
/model-config — show current primary + what's available
/model-config <model> — set the primary (e.g. minimax, claude, deepseek)
Steps
The brain is scripts/model_routing.py in claude-bootstrap. Resolve it via
~/.claude/.bootstrap-dir:
MR="$(cat ~/.claude/.bootstrap-dir)/scripts/model_routing.py"
1. Show current state (no argument)
echo "Available on this machine:"
python3 "$MR" detect | python3 -c "import sys,json;print(', '.join(k for k,v in json.load(sys.stdin).items() if v))"
echo "Current config:"
python3 "$MR" show
Report: which model is primary, the classifier, the mode (smart/hard),
and the available models. If auto_detected is true, mention it was inferred
from the machine and can be overridden.
2. Set a new primary (argument given)
python3 "$MR" set-primary "<model>" # validates against the detected set
python3 "$MR" apply # sync into srooter.yaml long_context
Then tell the user it takes effect for new sessions. If srooter is running, it must be restarted to pick up the routing change:
cd "$(cat ~/.claude/.bootstrap-dir)/../srooter" 2>/dev/null && \
echo "Restart srooter to apply: kill \$(cat .srooter.pid); and relaunch"
Notes
- Smart mode (default): the primary handles real coding; trivial/cheap asks still route to the local classifier (qwen) and explicit per-prompt overrides ("use claude") always win.
- No hardcoding — if the config is missing it is auto-created from what's installed/keyed on the machine.
Pre-analysis (MiniMax on every prompt)
When "analyze": true (default), the route-task hook sends each prompt to
MiniMax first for a terse INTENT / SCOPE / RISKS / APPROACH brief and injects
it into context, so Claude executes with that read of the task. Adds ~3-5s per
prompt. Toggle:
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.
- 8d ago First seen · 82 lines · 0 tokens per session scan B a604139d6846
model-config is a command published in the GitHub repository alinaqi/maggy (705 stars, last pushed 20d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 661 tokens. A static security scan graded it B 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 commands, from other repositories
crystallize
Transform verbose SOPs into high-performance agentic prompts via principled compression. Use when user says "/crystallize", "compress this prompt", "make this more agentic".
ai-feature-eval-harness
Design an evaluation plan for a product AI feature (LLM- or model-backed output): measurable success criteria, a held-out labeled eval dataset shape, per-criterion grading (code-based first, then LLM-based for nuanced judgment), and a pass threshold, then persist as AIEVALPLAN.md. Use when the task ships or changes a…
gentle-sdd-ff
Fast-forward all SDD planning phases — proposal through tasks.
sdd-explore
Explore and investigate an idea or feature — reads codebase and compares approaches.
sdd-init
Initialize SDD context — detects project stack and bootstraps persistence backend.
ijfw-audit
Run the IJFW audit gate for the current workflow phase. Usage: /ijfw-audit [phase name].