release: 0.6.7 — a curated example keeps its id, and an agent can say what it based a query on - #235
Merged
Merged
Conversation
… what it based a query on Two additions, and both exist to make something nameable that was not. ACE-109: `prompt_example.id` has been in the primary key since the table existed and was unusable — minted per deploy, never written back, and never selected by the read. So no example could be named, counted, or compared with itself across two deploys. The id is now derived from the example's own content, which is stable by construction and needs no coordination, and `example_by_id` reads one back. ACE-110: `execute_sql` takes an optional `basis` — what the agent based the query on, recorded beside the two self-reported columns the activity log already carries. The property is a bare array on purpose: the MCP SDK validates arguments before the handler runs, so a constraint in the schema refuses the whole call rather than trimming an entry, and the bound belongs at the boundary instead. Patch, following 0.6.3 and 0.6.4 — an `### Added` section is not a minor bump in this changelog's history. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
vishalkalbi27
requested review from
ashwin-agami and
sandeep-agami
as code owners
August 18, 2026 11:35
There was a problem hiding this comment.
Pull request overview
Prepares the 0.6.7 release by bumping declared package/plugin versions and documenting the two included feature changes (stable curated example IDs; optional basis recorded for execute_sql).
Changes:
- Bump version strings from
0.6.6to0.6.7in the plugin manifests and Python package metadata. - Add a
0.6.7entry toCHANGELOG.mddescribing ACE-109 and ACE-110.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| plugins/agami/.claude-plugin/plugin.json | Updates plugin version to 0.6.7. |
| packages/agami-core/pyproject.toml | Updates Python package version to 0.6.7. |
| CHANGELOG.md | Adds release notes for 0.6.7 dated 2026-08-18. |
| .claude-plugin/marketplace.json | Updates marketplace metadata/plugin version fields to 0.6.7. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
ashwin-agami
approved these changes
Aug 18, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Cuts 0.6.7 with the two changes merged since 0.6.6.
ACE-109 — a curated example keeps the same id across deploys.
prompt_example.idhas been in the primary key since the table existed and was unusable three ways over: minted per deploy when the YAML carried none, never written back, and never selected by the read. So nothing could name an example, count one, or tell whether the one returned yesterday is the one returned today. The id is now derived fromquestion+sql— stable by construction, no coordination, no write access to a customer's artifacts — andexample_by_idreads one back.ACE-110 — an agent can say what it based a query on. An optional
basisonexecute_sql, recorded ontool_callsbeside the two self-reported columns already there and rendered in the activity log under the statement it explains. The property is declared as a bare array deliberately: the MCP SDK validates arguments againstinputSchemabefore the handler runs, so a constraint there refuses the whole call rather than trimming one entry. Kinds are advertised in the description and enforced at the boundary.Version
Four files, the same set as 0.6.6 (
f1a4a01) —marketplace.jsoncarries it twice.Patch, not minor. 0.6.3 and 0.6.4 both shipped
### Addedsections as patches; this follows that. No compare link at the foot — that list stops at[0.3.4].Verification
pytest tests/— 4,643 passed, 12 skipped.One local failure,
test_org_resolver_defaults_to_local, is environmental and not from this branch: it fails identically on cleanmain, and passes with an emptyAGAMI_ARTIFACTS_DIR. The resolver was reading an org from the developer's own artifacts directory, which CI does not have.