token-pilot 0.23.6 → 0.23.7
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/.claude-plugin/plugin.json +1 -1
- package/CHANGELOG.md +22 -0
- package/dist/agents/tp-audit-scanner.md +1 -1
- package/dist/agents/tp-commit-writer.md +2 -1
- package/dist/agents/tp-dead-code-finder.md +1 -1
- package/dist/agents/tp-debugger.md +1 -1
- package/dist/agents/tp-history-explorer.md +1 -1
- package/dist/agents/tp-impact-analyzer.md +1 -1
- package/dist/agents/tp-migration-scout.md +1 -1
- package/dist/agents/tp-onboard.md +2 -1
- package/dist/agents/tp-pr-reviewer.md +1 -1
- package/dist/agents/tp-refactor-planner.md +1 -1
- package/dist/agents/tp-run.md +1 -1
- package/dist/agents/tp-session-restorer.md +2 -1
- package/dist/agents/tp-test-triage.md +1 -1
- package/dist/agents/tp-test-writer.md +1 -1
- package/package.json +1 -1
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "token-pilot",
|
|
3
|
-
"version": "0.23.
|
|
3
|
+
"version": "0.23.7",
|
|
4
4
|
"description": "Enforcement layer for token-efficient AI coding: MCP-first hook with structural denial summaries, SessionStart reminder, bless-agents CLI, and six tp-* subagents — works for every agent including those without MCP access.",
|
|
5
5
|
"author": "token-pilot",
|
|
6
6
|
"license": "MIT",
|
package/CHANGELOG.md
CHANGED
|
@@ -5,6 +5,28 @@ All notable changes to Token Pilot will be documented in this file.
|
|
|
5
5
|
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
|
|
6
6
|
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
|
|
7
7
|
|
|
8
|
+
## [0.23.7] - 2026-04-18
|
|
9
|
+
|
|
10
|
+
### Changed — per-agent `model:` selection for cheap, format-bound work
|
|
11
|
+
|
|
12
|
+
Claude Code allows each subagent to declare its own model in frontmatter (or `inherit` from the main agent). We've been relying on the user's global `CLAUDE_CODE_SUBAGENT_MODEL` env var as a blunt switch — that doesn't fit because some `tp-*` agents need real reasoning (debugger, impact analyzer, refactor planner) while others are pure format work. Moved three agents to **haiku-4.5** explicitly:
|
|
13
|
+
|
|
14
|
+
- **`tp-commit-writer`** — classifies diff → Conventional type, drafts short message. Context-bound, no architectural decisions.
|
|
15
|
+
- **`tp-session-restorer`** — parses `latest.md` + git status, emits a fixed-shape briefing. Pure transformation.
|
|
16
|
+
- **`tp-onboard`** — pulls project_overview and retells it in an orientation map. Format-bound.
|
|
17
|
+
|
|
18
|
+
The other 11 agents keep `inherit` — they do enough reasoning (intent, risk classification, call-tree traversal) that haiku would regress them. `tp-dead-code-finder` and `tp-audit-scanner` stay inherit for now; we'll revisit after real-world usage shows whether cross-check accuracy holds on haiku.
|
|
19
|
+
|
|
20
|
+
**User is NOT asked to set `CLAUDE_CODE_SUBAGENT_MODEL`.** The selection is per-agent and shipped with the template — predictable, rollback-friendly (one line per agent).
|
|
21
|
+
|
|
22
|
+
### Planned
|
|
23
|
+
|
|
24
|
+
- **TP-z64** (v0.28 backlog) — expanded tp-* roster with combo-agents that pair novel MCP-tool combinations for niche workflows (review-impact, test-coverage-gapper, api-surface-tracker, dep-health, incident-timeline). Must be brainstormed with names + triggers before implementation; deferred until v0.24 onboarding wizard ships and baseline stabilises.
|
|
25
|
+
- **v0.24.0** — onboarding wizard (doctor-warnings → one-step applied): writes `MAX_THINKING_TOKENS=10000` + `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE=50` to `~/.claude/settings.json`, generates `.claudeignore` if missing. Does NOT set `CLAUDE_CODE_SUBAGENT_MODEL` — per-agent model now handles that.
|
|
26
|
+
|
|
27
|
+
### Numbers
|
|
28
|
+
- 910 tests green, `tsc --noEmit` clean, 14 agents built.
|
|
29
|
+
|
|
8
30
|
## [0.23.6] - 2026-04-18
|
|
9
31
|
|
|
10
32
|
### Fixed — five findings from a live user audit
|
|
@@ -1,5 +1,6 @@
|
|
|
1
1
|
---
|
|
2
2
|
name: tp-commit-writer
|
|
3
|
+
model: claude-haiku-4-5-20251001
|
|
3
4
|
description: PROACTIVELY use this when the user is about to commit a NON-TRIVIAL change (new feature, fix, refactor) and asks "write a commit message". Reads staged diff, verifies tests pass, drafts Conventional Commit. Refuses mixed diffs (asks to split), failing tests, or empty stage. Do NOT use for docs-only, whitespace-only, or < 20-line diffs — the user can write those manually faster than a subagent spawn. Do NOT use to explain already-made commits.
|
|
4
5
|
tools:
|
|
5
6
|
- mcp__token-pilot__smart_diff
|
|
@@ -7,7 +8,7 @@ tools:
|
|
|
7
8
|
- mcp__token-pilot__test_summary
|
|
8
9
|
- mcp__token-pilot__outline
|
|
9
10
|
- Bash
|
|
10
|
-
token_pilot_version: "0.23.
|
|
11
|
+
token_pilot_version: "0.23.7"
|
|
11
12
|
token_pilot_body_hash: 559a0b61d20974bf33e35bc4c80dcf1b41d10d4df46cf9d05d3d5620713cd46f
|
|
12
13
|
---
|
|
13
14
|
|
|
@@ -1,5 +1,6 @@
|
|
|
1
1
|
---
|
|
2
2
|
name: tp-onboard
|
|
3
|
+
model: claude-haiku-4-5-20251001
|
|
3
4
|
description: PROACTIVELY use this when the user is exploring an unfamiliar codebase — asks "how is this organised", "what does this project do", "where do I start reading", or starts any conversation in a repo the main agent doesn't know. Orientation map only (layout, entry points, modules); does NOT drill into implementation.
|
|
4
5
|
tools:
|
|
5
6
|
- mcp__token-pilot__project_overview
|
|
@@ -9,7 +10,7 @@ tools:
|
|
|
9
10
|
- mcp__token-pilot__smart_read
|
|
10
11
|
- mcp__token-pilot__smart_read_many
|
|
11
12
|
- mcp__token-pilot__read_section
|
|
12
|
-
token_pilot_version: "0.23.
|
|
13
|
+
token_pilot_version: "0.23.7"
|
|
13
14
|
token_pilot_body_hash: ae0b86eaffaf34bf283b94b5572481fa8c2d6a2a25193f1173b70bef0fbe1919
|
|
14
15
|
---
|
|
15
16
|
|
|
@@ -7,7 +7,7 @@ tools:
|
|
|
7
7
|
- mcp__token-pilot__read_diff
|
|
8
8
|
- mcp__token-pilot__outline
|
|
9
9
|
- mcp__token-pilot__read_symbol
|
|
10
|
-
token_pilot_version: "0.23.
|
|
10
|
+
token_pilot_version: "0.23.7"
|
|
11
11
|
token_pilot_body_hash: a058518619fd6e2def0c9226f6c70438a5e0a80efe680c935414ecd7e1b14a4f
|
|
12
12
|
---
|
|
13
13
|
|
package/dist/agents/tp-run.md
CHANGED
|
@@ -1,5 +1,6 @@
|
|
|
1
1
|
---
|
|
2
2
|
name: tp-session-restorer
|
|
3
|
+
model: claude-haiku-4-5-20251001
|
|
3
4
|
description: PROACTIVELY use this as the FIRST step after /clear, compaction, or a fresh window when a recent session_snapshot exists on disk. Reads snapshot + git status + saved docs, returns a ≤200-token briefing. Do NOT use mid-task.
|
|
4
5
|
tools:
|
|
5
6
|
- mcp__token-pilot__smart_read
|
|
@@ -8,7 +9,7 @@ tools:
|
|
|
8
9
|
- mcp__token-pilot__session_budget
|
|
9
10
|
- Bash
|
|
10
11
|
- Read
|
|
11
|
-
token_pilot_version: "0.23.
|
|
12
|
+
token_pilot_version: "0.23.7"
|
|
12
13
|
token_pilot_body_hash: 35b7f333a28c94e7dc89fcc3171703c4b466225f55cd5c701b7592f4f6486440
|
|
13
14
|
---
|
|
14
15
|
|
|
@@ -7,7 +7,7 @@ tools:
|
|
|
7
7
|
- mcp__token-pilot__read_range
|
|
8
8
|
- mcp__token-pilot__find_usages
|
|
9
9
|
- mcp__token-pilot__read_symbol
|
|
10
|
-
token_pilot_version: "0.23.
|
|
10
|
+
token_pilot_version: "0.23.7"
|
|
11
11
|
token_pilot_body_hash: 255912c47661d203c8f9a735237bc419f97e937f788a01811bbe126ee3dd5878
|
|
12
12
|
---
|
|
13
13
|
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "token-pilot",
|
|
3
|
-
"version": "0.23.
|
|
3
|
+
"version": "0.23.7",
|
|
4
4
|
"description": "Save up to 80% tokens when AI reads code — MCP server for token-efficient code navigation, AST-aware structural reading instead of dumping full files into context window",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"main": "dist/index.js",
|