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 skills/aksoftcode/aicrew/aicrew-benchmarknpx skills add AKSoftCode/aicrew --skill aicrew-benchmarkgit clone --depth 1 https://github.com/AKSoftCode/aicrewWhat 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 | $0.00026 | $0.00581 |
| Opus 5 | $0.00013 | $0.00291 |
| Sonnet 5 | $0.00005 | $0.00116 |
| Haiku 4.5 | $0.00003 | $0.00058 |
Grade A, and why
aicrew-benchmark 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 2d 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 — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
aicrew-benchmark (Codex)
Use when: measuring aicrew skill output quality, latency, or token usage across skills or pipelines.
Equivalent actions
| Method | Command |
|---|---|
| CLI | aicrew benchmark (planned) |
| Codex skill | aicrew-benchmark (this skill) |
| Claude Code slash | /benchmark |
Default output
Caveman/lean style. See ~/Agents/agents/caveman.md.
Source of truth:
~/Agents/commands/harness-audit.md(for harness health)
What to benchmark
Ask the user which dimension to measure:
What should we benchmark?
- Skill trigger accuracy — does the right skill fire for a given prompt?
- Output quality — does a skill produce the expected sections/format?
- Token efficiency — lean vs normal mode token counts for the same task
- Pipeline timing — wall-clock time for each /dev or /fix phase
- All of the above (full harness benchmark)
Wait for answer.
Steps per benchmark type
1. Skill trigger accuracy
- List all skills and their
descriptionfields - For each skill, present a sample prompt and confirm the right skill triggers
- Flag any ambiguous descriptions
2. Output quality
- Run a skill against a standard input
- Check required sections are present (e.g. SCOUT: block, HANDOFF: block, phase gates)
- Flag missing or malformed sections
3. Token efficiency
- Compare output token count with lean vs normal mode
- Run the same task in both modes; measure diff
4. Pipeline timing
- Record start/end of each phase in
/devor/fix - Output phase timings as a table
5. Full harness
- Run all checks above in sequence
- Output a summary table: skill → trigger ✓/✗ → quality ✓/✗ → tokens → timing
Output format
BENCHMARK:
Skill/Phase | Trigger | Quality | Tokens | Timing
--------------|---------|---------|--------|-------
aicrew-dev | ✓ | ✓ | 1200 | 4.2s
aicrew-fix | ✓ | ✓ | 800 | 2.1s
...
Checkpoint
What ships with it
1 file 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.
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.
- 2d ago First seen · 76 lines · 26 tokens per session scan A cc8b868523c1
aicrew-benchmark is a skill published in the GitHub repository AKSoftCode/aicrew (3 stars, last pushed 2mo ago), licensed MIT. It adds 26 tokens to every session and 581 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 2026-08-31.
Other skills, from other repositories
sdd
Execute the Liatrio Spec-Driven Development (SDD) workflow when explicitly invoked by the user. NOTE: this skill is NOT intended to be dynamically loaded or automatically triggered; it should only ever be explicitly called by the user.
repo-layout
Register, resolve, and validate where a project's path-slots actually live via .layout-slots.yaml markers, so relocating a slot needs zero SKILL.md edits. Do NOT use for single-file lookups.
compiling-project-guidelines
Compile scattered guideline sources into one scope-aware COMPILED-GUIDELINES.md for other skills and ult-context-generate's Constraints layer. Do NOT use to enforce rules at runtime.
vendored-skill
Vendored reference copy, for reproducibility of the consumer-benefit-user-stories case study. NOT part of this repo's installable skill set — origin: ground-up, maintained outside context-engineering-oss, in the same spirit as the FastAPI/Textual/Open5GS corpora this repo is demonstrated against rather than adopting.…
cep-wizard
Launch a local, localhost-only browser wizard with two journeys — layout onboarding for a project with ult-repo-layout installed, initialized or not — guides an uninitialized or not-yet-discovered repo through Run Discover, then shows resolved layer/slot state through four labeled boxes and a directory picker, and…
codegraph
Generate a codebase knowledge graph with graphify at graphify-out/ so other skills can query cross-file relationships before touching code. Do NOT use for runtime profiling.