Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions .claude/skills/bat-story-eval/SKILL.md
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,7 @@ If `$ARGUMENTS` is `--help` or missing `--baseline`, show usage and stop:
### 0a. Compute Diff

```bash
cd /home/julien/github/ha-mcp/worktree/uat-stories
cd "$(dirname "$(git rev-parse --path-format=absolute --git-common-dir)")/worktree/uat-stories"
git diff <baseline>..HEAD -- src/ha_mcp/ --stat
git diff <baseline>..HEAD -- src/ha_mcp/ --name-only
```
Expand Down Expand Up @@ -112,7 +112,7 @@ For EACH agent, run all stories against the **baseline** version. One container
### 1a. Start container with first story

```bash
cd /home/julien/github/ha-mcp/worktree/uat-stories
cd "$(dirname "$(git rev-parse --path-format=absolute --git-common-dir)")/worktree/uat-stories"
uv run python tests/uat/stories/run_story.py \
catalog/<first_story>.yaml \
--agents <agent> --keep-container \
Expand Down Expand Up @@ -370,4 +370,4 @@ Flag >5% total size increase (directly impacts token cost per turn).
- Reuse containers: first story starts it (`--keep-container`), rest use `--ha-url`
- Custom story YAMLs go to `/tmp/` (ephemeral); full details reported in Step 6
- See "Metrics" section in Step 4 for primary vs secondary metric classification
- The working directory MUST be `/home/julien/github/ha-mcp/worktree/uat-stories` for `uv run`
- The working directory MUST be the `worktree/uat-stories` worktree root (where `pyproject.toml` lives) for `uv run`
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@ uv run python tests/uat/stories/scripts/ha_query.py \
Check the exit code first. `ha_query.py` exits non-zero when the agent CLI
itself failed, and prints `[exit N]` (plus stderr, when there is any) after
whatever text the CLI managed to produce — `[exit 124]` for a query that hung
past its timeout. The one non-zero exit without a marker is `Error: <agent>
past its timeout. The one non-zero exit without a marker is `Error: {agent}
CLI not found`, which means the agent is not installed: fix the environment
rather than re-running. A failed query is **not a measurement**: do not score
it as any of the three outcomes below — re-run it, and if it keeps failing,
Expand Down
Loading