Skip to content

Attocode Intelligence 0.1.8

Choose a tag to compare

View all tags
@eren23 eren23 released this 17 Oct 14:47
· 118 commits to main since this release
intel-v0.1.8
94d345b
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Attocode Intelligence 0.1.8 adds exact file/symbol selection and smaller changed-class summaries.

uv tool install --force attocode-code-intel==0.1.8
attocode-code-intel changes --file src/worker.py --symbol Worker.work
attocode-code-intel changes --class-mode expand

Repeat file/symbol selectors to focus the report before caps. Refresh retains the original selection and class mode; saved reports still detect changes across the complete Git snapshot.

Changed classes now default to bounded identity/member summaries. Changed methods retain their own caller paths. Class-wide callers are deliberately omitted by default: use --class-mode expand for the previous traversal, or trace a class directly. Full local MCP supports the same selectors and class modes. Daily keeps its 16 tools, existing client configurations are preserved, and ranking defaults stay unchanged. Restart MCP after upgrading.

Three historical Python replays preserved 45 non-class caller flows; class-heavy reports were 32.5% and 47.9% smaller. This does not establish better agent outcomes or general speed.

All 17 release jobs passed, including Python 3.12/3.13 product and engine tests, PostgreSQL, the combined suite, four real language servers, ranking, Linux/macOS wheel checks, Docker and dashboard. Uploaded XML verified all 22 focus, 29 existing workflow and 51 SQL/HTTP cases without skips on both Python versions, including PostgreSQL locking controls. Actual PyPI clean installation and upgrade from 0.1.7 passed all 11 checks, including daily/lean/full MCP, CLI filters/summary/expand/export/check/refresh, persisted index and three-client configuration preservation. Every packaged Python file byte-matches the tagged source.

Attached wheel and source archive were downloaded from PyPI and checked against its SHA-256 metadata. The verification JSON records artifact hashes, exact source, release workflow, test counts, install/upgrade timings and temporary-storage cleanup. No new cloud benchmark or model study ran.

Release notes · Workflow guide · PyPI

Assets 5
Loading