feat(memory): surface active decisions at session start + capture nudge (Context Recovery)

Context Recovery now shows recent scope-relevant active decisions (bounded read of
decisions.active.json via gstack-decision-search) and instructs the agent to treat
them as settled calls and to log durable decisions/reversals. Closes the Phase-1
capture->curate->resurface loop, reliable + file-only. Regen across all hosts folded
in (squash-with-regen); parity 10/10, freshness green.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
Garry Tan
2026-06-07 09:53:21 -07:00
parent a52e033576
commit c375788e6c
46 changed files with 309 additions and 35 deletions
-8
View File
@@ -22,7 +22,6 @@ triggers:
<!-- AUTO-GENERATED from SKILL.md.tmpl — do not edit directly -->
<!-- Regenerate: bun run gen:skill-docs -->
## When to invoke this skill
Use when asked to "ship", "deploy",
@@ -404,7 +403,6 @@ Before calling AskUserQuestion, verify:
- [ ] If you split, you checked dependencies between options before firing the chain
- [ ] If a per-option Hold fires, you stopped the chain immediately (didn't queue)
## Artifacts Sync (skill start)
```bash
@@ -502,8 +500,6 @@ else
fi
```
Privacy stop-gate: if output shows `ARTIFACTS_SYNC: off`, `artifacts_sync_mode_prompted` is `false`, and gbrain is on PATH or `gbrain doctor --fast --json` works, ask once:
> gstack can publish your artifacts (CEO plans, designs, reports) to a private GitHub repo that GBrain indexes across machines. How much should sync?
@@ -530,7 +526,6 @@ At skill END before telemetry:
"~/.claude/skills/gstack/bin/gstack-brain-sync" --once 2>/dev/null || true
```
## Model-Specific Behavioral Patch (claude)
The following nudges are tuned for the claude model family. They are
@@ -604,7 +599,6 @@ Applies to AskUserQuestion, user replies, and findings. AskUserQuestion Format i
Curated jargon list lives at `~/.claude/skills/gstack/scripts/jargon-list.json` (80+ terms). On the first jargon term you encounter this session, Read that file once; treat the `terms` array as the canonical list. The list is repo-owned and may grow between releases.
## Completeness Principle — Boil the Lake
AI makes completeness cheap. Recommend complete lakes (tests, edge cases, error paths); flag oceans (rewrites, multi-quarter migrations).
@@ -780,8 +774,6 @@ branch name wherever the instructions say "the base branch" or `<default>`.
---
# Ship: Fully Automated Ship Workflow
You are running the `/ship` workflow. This is a **non-interactive, fully automated** workflow. Do NOT ask for confirmation at any step. The user said `/ship` which means DO IT. Run straight through and output the PR URL at the end.
-15
View File
@@ -390,7 +390,6 @@ Before calling AskUserQuestion, verify:
- [ ] If you split, you checked dependencies between options before firing the chain
- [ ] If a per-option Hold fires, you stopped the chain immediately (didn't queue)
## Artifacts Sync (skill start)
```bash
@@ -488,8 +487,6 @@ else
fi
```
Privacy stop-gate: if output shows `ARTIFACTS_SYNC: off`, `artifacts_sync_mode_prompted` is `false`, and gbrain is on PATH or `gbrain doctor --fast --json` works, ask once:
> gstack can publish your artifacts (CEO plans, designs, reports) to a private GitHub repo that GBrain indexes across machines. How much should sync?
@@ -516,7 +513,6 @@ At skill END before telemetry:
"$GSTACK_BIN/gstack-brain-sync" --once 2>/dev/null || true
```
## Model-Specific Behavioral Patch (claude)
The following nudges are tuned for the claude model family. They are
@@ -590,7 +586,6 @@ Applies to AskUserQuestion, user replies, and findings. AskUserQuestion Format i
Curated jargon list lives at `$GSTACK_ROOT/scripts/jargon-list.json` (80+ terms). On the first jargon term you encounter this session, Read that file once; treat the `terms` array as the canonical list. The list is repo-owned and may grow between releases.
## Completeness Principle — Boil the Lake
AI makes completeness cheap. Recommend complete lakes (tests, edge cases, error paths); flag oceans (rewrites, multi-quarter migrations).
@@ -766,8 +761,6 @@ branch name wherever the instructions say "the base branch" or `<default>`.
---
# Ship: Fully Automated Ship Workflow
You are running the `/ship` workflow. This is a **non-interactive, fully automated** workflow. Do NOT ask for confirmation at any step. The user said `/ship` which means DO IT. Run straight through and output the PR URL at the end.
@@ -807,8 +800,6 @@ Never skip a verification step because a prior `/ship` run already performed it.
---
---
## Step 1: Pre-flight
@@ -1937,8 +1928,6 @@ Substitute: TIMESTAMP = ISO 8601 datetime, STATUS = "clean" if 0 findings or "is
Include any design findings alongside the code review findings. They follow the same Fix-First flow below.
### Step 9.3: Cross-review finding dedup
Before classifying findings, check if any were previously skipped by the user in a prior review on this branch.
@@ -2057,8 +2046,6 @@ For each comment in `comments`:
---
## Capture Learnings
If you discovered a non-obvious pattern, pitfall, or architectural insight during
@@ -2084,8 +2071,6 @@ staleness detection: if those files are later deleted, the learning can be flagg
**Only log genuine discoveries.** Don't log obvious things. Don't log things the user
already knows. A good test: would this insight save time in a future session? If yes, log it.
### Refresh learnings for the headline feature on this branch
The top-of-skill learnings pull was keyed to "release ship" broadly. Before the VERSION/CHANGELOG step, re-pull learnings keyed to THIS branch's headline feature so any prior version-bump or CHANGELOG pitfalls for similar features surface.
-11
View File
@@ -392,7 +392,6 @@ Before calling AskUserQuestion, verify:
- [ ] If you split, you checked dependencies between options before firing the chain
- [ ] If a per-option Hold fires, you stopped the chain immediately (didn't queue)
## Artifacts Sync (skill start)
```bash
@@ -490,8 +489,6 @@ else
fi
```
Privacy stop-gate: if output shows `ARTIFACTS_SYNC: off`, `artifacts_sync_mode_prompted` is `false`, and gbrain is on PATH or `gbrain doctor --fast --json` works, ask once:
> gstack can publish your artifacts (CEO plans, designs, reports) to a private GitHub repo that GBrain indexes across machines. How much should sync?
@@ -518,7 +515,6 @@ At skill END before telemetry:
"$GSTACK_BIN/gstack-brain-sync" --once 2>/dev/null || true
```
## Model-Specific Behavioral Patch (claude)
The following nudges are tuned for the claude model family. They are
@@ -592,7 +588,6 @@ Applies to AskUserQuestion, user replies, and findings. AskUserQuestion Format i
Curated jargon list lives at `$GSTACK_ROOT/scripts/jargon-list.json` (80+ terms). On the first jargon term you encounter this session, Read that file once; treat the `terms` array as the canonical list. The list is repo-owned and may grow between releases.
## Completeness Principle — Boil the Lake
AI makes completeness cheap. Recommend complete lakes (tests, edge cases, error paths); flag oceans (rewrites, multi-quarter migrations).
@@ -768,8 +763,6 @@ branch name wherever the instructions say "the base branch" or `<default>`.
---
# Ship: Fully Automated Ship Workflow
You are running the `/ship` workflow. This is a **non-interactive, fully automated** workflow. Do NOT ask for confirmation at any step. The user said `/ship` which means DO IT. Run straight through and output the PR URL at the end.
@@ -809,8 +802,6 @@ Never skip a verification step because a prior `/ship` run already performed it.
---
---
## Step 1: Pre-flight
@@ -2462,8 +2453,6 @@ staleness detection: if those files are later deleted, the learning can be flagg
**Only log genuine discoveries.** Don't log obvious things. Don't log things the user
already knows. A good test: would this insight save time in a future session? If yes, log it.
### Refresh learnings for the headline feature on this branch
The top-of-skill learnings pull was keyed to "release ship" broadly. Before the VERSION/CHANGELOG step, re-pull learnings keyed to THIS branch's headline feature so any prior version-bump or CHANGELOG pitfalls for similar features surface.