Borrowing it
Nothing to install: this file belongs to rudraptpsingh/axon. 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/rudraptpsingh/axon/master/.claude/commands/eval-axon.mdgit clone --depth 1 https://github.com/rudraptpsingh/axonWrote 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/rudraptpsingh/axon/eval-axon)<a href="https://agentmods.dev/commands/rudraptpsingh/axon/eval-axon"><img src="https://agentmods.dev/badge/commands/rudraptpsingh/axon/eval-axon/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/commands/rudraptpsingh/axon/eval-axon"><img src="https://agentmods.dev/badge/commands/rudraptpsingh/axon/eval-axon.svg" alt="Reviewed on agentmods" width="80" 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 | $0.01479 |
| Opus 5 | $0.00000 | $0.00740 |
| Sonnet 5 | $0.00000 | $0.00296 |
| Haiku 4.5 | $0.00000 | $0.00148 |
Grade A, and why
eval-axon 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 9d 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are evaluating axon — a zero-cloud MCP server that gives AI coding agents real-time hardware awareness. You ARE the agent. axon is YOUR tool.
Setup
- Build and install axon:
cargo install --path crates/axon-cli - Run
axon setup claude-code - Verify:
axon diagnoseandaxon query hw_snapshot
Evaluation Loop (repeat 5 times minimum)
For each iteration:
Step 1: Do real work while consulting axon
- Before heavy tasks (cargo build, cargo test, file writes): run
axon query hw_snapshot— check headroom field. If insufficient, reduce parallelism or defer. - During stress: run
axon query process_blame— check if culprit/group/fix are accurate and actionable. - After stress: run
axon diagnose— check if recovery is detected, if recent alerts show. - End of session: run
axon query session_health— check if the narrative captures what happened.
Step 2: Generate stress to test detection
- CPU stress:
for i in 1 2 3 4; do yes > /dev/null & donethen observe axon - Memory stress:
python3 -c "x=[bytearray(100_000_000) for _ in range(50)]"then observe - Mixed: run cargo clean && cargo build --release -j 4 with yes workers simultaneously
- Kill stress and immediately diagnose — check recovery indicator
- I/O stress:
dd if=/dev/zero of=/tmp/axon_testfile bs=1M count=4096 conv=fdatasyncthen observe — check if impact score reflects disk saturation - Gradual leak:
python3 -c "import time; x=[]; [x.append(bytearray(50_000_000)) or time.sleep(2) for _ in range(20)]"— watch for slow drift detection over ~40s - Flap test: rapidly alternate stress and idle (5s on, 5s off, repeat 6 times) — verify no alert storm, hysteresis prevents flapping
- Recovery test: start stress, wait 10s, kill stress, verify recovery/resolved alert fires within 10s
- Alert verification: after stress, run
axon query session_health-- check alert_count > 0; if 0, alerts are not firing when they should - Serve lifecycle: start
axon servevia MCP stdio, run all 7 tools, verify serve process stays alive throughout, verify clean exit on stdin close - GPU snapshot: run
axon query gpu_snapshot-- verify ok=true and detected field is present
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.
- 9d ago First seen · 100 lines · 0 tokens per session scan A ff210773f2f5
eval-axon is a command published in the GitHub repository rudraptpsingh/axon (9 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,479 tokens. 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 2026-08-31.
Other commands, from other repositories
init
Create or update the AgentBridge project config files in the current workspace.
ng-context-pack
Portable command prompt generated from skills/briefing-an-agent/SKILL.md. Edit the skill, then run python tools/ng.py gen-commands; do not edit this file by hand.
ng-red-team
Portable command prompt generated from skills/stress-testing-agent-changes/SKILL.md. Edit the skill, then run python tools/ng.py gen-commands; do not edit this file by hand.
task-verify
PitWay: Run an inprogress task's approved verification command and persist evidence.
verification-repair
PitWay: Bounded, approve-before-edit correction after every task is done.
usage-add
PitWay: Accumulate measured planning or qa token usage onto a milestone.