Skip to content

fix: project health follow-ups, cost planner label, altimate-dbt docs - #41

Merged
hkaltai merged 4 commits into
mainfrom
fix/plugin-follow-ups
Nov 20, 2026
Merged

hkaltai merged 4 commits into
mainfrom
fix/plugin-follow-ups

Conversation

@hkaltai

@hkaltai hkaltai commented Nov 19, 2026

Copy link
Copy Markdown
Contributor

Follow-ups from the merged project health (#33) and scan-cost (#39) reviews, plus a docs section.

Changes

  • Project health: the start folder of a Bash command with its own cd is kept per call (tool_use_id), so an overlapping call (a subagent's) no longer loses it and leaves a written model unchecked.
  • Project health: echo ${#arr[@]}; git push is read as a push (the # of ${#name} is not a comment).
  • Project health: the Bash gate and after-phase return at once in a session with no dbt project enabled (no shell-folder lookup, no hold for a push).
  • Tests for the start-folder rules and the widening after a baseline by a YAML file no insight names.
  • /scan-cost with no model keeps the planner label of the last Snowflake or Databricks estimate.
  • README: a short altimate-dbt section (what it is, what the agent uses it for, that users need not run it, link to the docs).
  • Version 1.8.3 and changelog entry (1.8.2 is taken by the open usage-tracking PR).

Not changed

  • The state.running || clause in the push decision cannot be observed on its own: the chain's outer bound reports first, so no test can tell it apart.
  • The fallback to the session root for a cd command with no noted start folder is unreachable without outlasting the chain deadline; the null-folder path to the same root is tested.

Test plan

  • claude plugin test plugins/altimate-code (1191 pass; the known artifact-read race test is the only failure)
  • claude plugin validate --strict .
  • bash tests/*.sh

🤖 Generated with Claude Code

…led-session gate), cost planner label kept, altimate-dbt docs

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@hkaltai
hkaltai marked this pull request as ready for review November 19, 2026 14:28
// What is shown never decides the gate: a failure here must not let a scan the gate knows is over the limit through.
try {
await update($, costAtom, prev => ({ ...(prev ?? FRESH), lastEstimate: { target, bytes, display } }))
await update($, costAtom, prev => ({ ...(prev ?? FRESH), lastEstimate: { target, bytes, display, ...(planner && { planner }) } }))

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

SUGGESTION: planner is now written into lastEstimate, but its declared shape was not updated.

CostState.lastEstimate in types/index.d.ts:507 is still { target: string; bytes: number; display: string }, so the persisted state and its documented type disagree (the sibling held field already declares planner?: string). Reading last.planner anywhere later would fail typecheck, and the type comment no longer describes what is stored. Add planner?: string there (the spread hides the excess property from typecheck here).

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed in 2a08944: planner?: string is declared on CostState.lastEstimate.


import { PHASE_TIMEOUT_MS } from './tool-chain'
import { GATE_WAIT_MS, bandText, canonical, configPathProblem, filesWrittenBy, isGitPush, modelsOfFiles, newIssues, pairKey, pushRefusal, reportText, simpleCommands, toolArgs } from './features/health'
import { GATE_WAIT_MS, bandText, bashAfter, canonical, configPathProblem, filesWrittenBy, isGitPush, modelsOfFiles, newIssues, pairKey, pushRefusal, reportText, simpleCommands, toolArgs } from './features/health'

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

SUGGESTION: bashAfter is imported here but never used.

The new tests (lines 868-913) exercise the after-phase through $.tool.call(...), not by calling bashAfter directly, so this name is dead in the import list. Drop it so the import keeps matching what the file uses.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed in df2497a: the unused import is removed.

@kilo-code-bot

kilo-code-bot Bot commented Nov 19, 2026 •

Copy link
Copy Markdown

Code Review Summary

Status: No Issues Found | Recommendation: Merge

The incremental change removes the unused bashAfter import from health.test.tsx; its only remaining occurrences are gone and every other imported name is still used. The earlier CostState.lastEstimate finding remains fixed at types/index.d.ts:507 (planner?: string), consistent with cost.tsx:769.

Files Reviewed (1 file)
  • plugins/altimate-code/hooks/health.test.tsx - no issues (unused import removed; all remaining imports used)

Fix these issues in Kilo Cloud

Previous Review Summaries (2 snapshots, latest commit 2a08944)

Current summary above is authoritative. Previous snapshots are kept for context only.

Previous review (commit 2a08944)

Status: 1 Issue Found | Recommendation: Address before merge

Overview

Severity Count
CRITICAL 0
WARNING 0
SUGGESTION 1

The incremental change is a one-line fix: CostState.lastEstimate now declares planner?: string, matching what rememberEstimate stores at cost.tsx:769 (...(planner && { planner }) only spreads a non-empty string, so the optional field is accurate and consistent with the sibling held field). That resolves the earlier lastEstimate type mismatch.

Issue Details (click to expand)

SUGGESTION

File Line Issue
plugins/altimate-code/hooks/health.test.tsx 4 bashAfter is still imported but never used (its only occurrence is the import line); the after-phase tests go through $.tool.call(...). Re-verified at current HEAD 2a08944.
Files Reviewed (2 files)
  • plugins/altimate-code/types/index.d.ts - no issues (fix verified against cost.tsx:769)
  • plugins/altimate-code/hooks/health.test.tsx - 1 issue (targeted re-check, unchanged file)

Fix these issues in Kilo Cloud

Previous review (commit 7f43f88)

Status: 2 Issues Found | Recommendation: Address before merge

Overview

Severity Count
CRITICAL 0
WARNING 0
SUGGESTION 2

The functional changes check out: the ${#name} guard is needed ({/} are in OPERATORS, so { really does end the word and without the guard the ; git push becomes a comment), the per-call start-folder map is correctly keyed by tool_use_id (the same e flows from bashGate to bashAfter in tool-chain.ts), its 50-entry cap is bounded, a refused call's stale note is always overwritten by noteFolder before any later read, and the enabled early returns in bashGate/bashAfter match markChanged's existing projectAbove(root) guard. The cost gate//scan-cost label propagation and the version/changelog/marketplace bumps are consistent.

Issue Details (click to expand)

SUGGESTION

File Line Issue
plugins/altimate-code/hooks/features/cost.tsx 769 planner is now persisted in lastEstimate, but CostState.lastEstimate in types/index.d.ts:507 still declares only { target, bytes, display }; add planner?: string (as held has) so the state's documented type matches what is stored.
plugins/altimate-code/hooks/health.test.tsx 4 bashAfter was added to the import list but is never used; the new tests exercise the after-phase through $.tool.call(...).

Fix these issues in Kilo Cloud

Files Reviewed (8 files)
  • .claude-plugin/marketplace.json - no issues
  • plugins/altimate-code/.claude-plugin/plugin.json - no issues
  • plugins/altimate-code/CHANGELOG.md - no issues
  • plugins/altimate-code/README.md - no issues
  • plugins/altimate-code/hooks/features/cost.tsx - 1 issue
  • plugins/altimate-code/hooks/features/cost.test.tsx - no issues
  • plugins/altimate-code/hooks/features/health.tsx - no issues
  • plugins/altimate-code/hooks/health.test.tsx - 1 issue

Reviewed by deepseek-v4.1-flash-1017 · Input: 0 · Output: 0 · Cached: 0

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@hkaltai

hkaltai commented Nov 19, 2026

Copy link
Copy Markdown
Contributor Author

review-loop: APPROVED at 2a08944 (reviewers: gpt-6-astra high, gpt-6.1-sol high)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@hkaltai

hkaltai commented Nov 19, 2026

Copy link
Copy Markdown
Contributor Author

review-loop: APPROVED at df2497a (reviewers: gpt-6-astra high, gpt-6.1-sol high)

…ng 1.8.2)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@hkaltai

hkaltai commented Nov 19, 2026

Copy link
Copy Markdown
Contributor Author

review-loop: APPROVED at 8c389fc (reviewers: gpt-6-astra high, gpt-6.1-sol high)

@saravmajestic saravmajestic left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. The per-call start folder (tool_use_id, newest 50 kept), the ${#name} comment fix, and the early return when no dbt project is enabled all look right, and both Kilo findings are fixed. Locally: 1231 pass, 0 fail, and 1264/0 merged with #42 on main. The two conflict only in CHANGELOG.md.

@hkaltai
hkaltai merged commit 660fd09 into main Nov 20, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants