Skip to content

Agent skill

The aster agent skill is the agent-facing form of the bench. It is deliberately thin: it teaches an agent when to use the bench, where the APIs are, and which evidence rules must survive the loop.

What is in the skill

skills/aster/
├── SKILL.md                 # trigger conditions, loop, non-negotiables
├── references/
│   ├── api-map.md           # task → API index
│   ├── aedt-contract.md     # guard/AEDT configuration + run lifecycle
│   └── evidence-rules.md    # labels, kept artifacts, re-scoring
└── scripts/
    └── dry_demo.py          # runnable starter, no AEDT

Install

With the library installed:

aster skill install            # → %USERPROFILE%\.dsh\skills\aster
aster skill path               # print that destination

Without installing, copy the folder into the agent's skill directory:

Copy-Item -Recurse skills\aster "$env:USERPROFILE\.dsh\skills\aster"

For other agent platforms, keep the same layout (SKILL.md + references/ + scripts/) and point the agent at the folder. The YAML front matter carries the skill name and description; the body is the instruction.

Try it

python skills\aster\scripts\dry_demo.py --samples 32

The script exercises the full dry loop: parameter space → Latin hypercube → evaluations → Pareto archive.

What the skill will not do

  • It will not present a PREDICTED score as measured.
  • It will not drop a failed run silently.
  • It will not drive the web GUI (not published yet).