SDK AI Bot Run Evaluation

What it does

When your agent uses it

• Uploading datasets
• Pipeline troubleshooting
• Knowledge-graph indexing

Example prompts

• “run evaluation”
• “run eval”
• “evaluate the bot”
• “/sdk-ai-bot-run-evaluation”

Requirements

• Python 3
• A credential in BOT_AGENT_ACCESS_TOKEN
• Compatibility (from SKILL.md): local azure-sdk-tools clone, python 3.12 venv, az login, bot /completion endpoint

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Ensure the required env vars are set; for a local run, start the agent server.py.
2. 2Choose the dataset: a scenario file, an asset name, or a one-row file for a single case.
3. 3Run python evals_run.py --dataset <...> --is_ci False (add --baseline_check False
4. 4Open the printed Foundry Report URL and review the per-case score table.
5. 5With --cache_result full, inspect cache/-*.json for per-case and

What it can do on your machine

Read from SKILL.md and the folder at commit 942ef24. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • python
  • az

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md. Its commands use az, which can reach the network depending on how they are called.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names these keys or tokens, usually read from environment variables:

  • BOT_AGENT_ACCESS_TOKEN

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

• Compatibility

  local azure-sdk-tools clone, python 3.12 venv, az login, bot /completion endpoint

  From compatibility in the SKILL.md frontmatter.

Context cost

SDK AI Bot Run Evaluation loads about 1.1k tokens when it runs, and up to ~2.3k if it reads all its reference files. Until then it costs about 108 tokens; SKILL.md has 386 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~108

When it runs · the whole SKILL.md, loaded when a task matches

~1.1k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~2.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

• NoteMentions a .env fileSKILL.md:36

  user to set these before running (loads `.env`; copy and fill

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from Azure/azure-sdk-tools at commit 942ef24, republished under its MIT licence (© Azure). 386 words, ~1,148 tokens.

name

sdk-ai-bot-run-evaluation

description

Run Azure SDK QA bot evaluations on curated datasets locally, including a single test case. WHEN: "run evaluation", "run eval", "evaluate the bot", "run perf evaluation", "run basic evaluation", "run a single test case", "evaluate one question", "run all scenarios", "score the bot", "run evals locally". DO NOT USE FOR: preparing or uploading datasets, pipeline troubleshooting, knowledge-graph indexing.

compatibility

local azure-sdk-tools clone, python 3.12 venv, az login, bot /completion endpoint

license

MIT

metadata.version

1.0.0

metadata.distribution

local

# One scenario (all evaluators), against the local server, no baseline gate:
python evals_run.py --dataset evaluation_datasets/perf/typespec.jsonl \
  --is_ci False --baseline_check False --cache_result full

# Same via the asset name:
python evals_run.py --dataset "qa-bot-perf-typespec:latest" --is_ci False --baseline_check False

# Subset of evaluators:
python evals_run.py --dataset "qa-bot-basic-python:latest" \
  --evaluators "bot_evals,groundedness" --is_ci False --baseline_check False

© Azure, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (references) in .github/skills/sdk-ai-bot-run-evaluation of Azure/azure-sdk-tools.

• SKILL.md
• references/running-evaluations.md

Open the folder on GitHubat commit 942ef24