uv-suite 0.12.0 → 0.14.0

package/hooks/session-end.sh

@@ -36,10 +36,10 @@ if [ -n "$STAGED" ] || [ -n "$UNSTAGED" ] || [ -n "$UNTRACKED" ]; then
   REVIEW_MSG="Uncommitted changes — consider /review and /slop-check before committing. "
 fi
 
-# Reflection prompt
-REFLECTION_MSG="Before closing: What shipped? What did you learn? What would you teach the agent for next time (add to CLAUDE.md or DANGER-ZONES.md)?"
+# Checkpoint prompt
+CHECKPOINT_MSG="Run /checkpoint to save session state for next time. Run /restore at the start of your next session."
 
-FULL_MSG="${DURATION_MSG}${REVIEW_MSG}${REFLECTION_MSG}"
+FULL_MSG="${DURATION_MSG}${REVIEW_MSG}${CHECKPOINT_MSG}"
 
 cat <<EOF
 {

package/install.sh

@@ -208,7 +208,7 @@ This project uses [UV Suite](https://github.com/utsavanand/uv-suite) v${UV_VERSI
 
 ### Skills
 
-/map-codebase, /map-stack, /spec, /architect, /review, /write-tests, /write-evals, /slop-check, /prototype, /security-review
+/map-codebase, /map-stack, /spec, /architect, /review, /write-tests, /write-evals, /slop-check, /prototype, /security-review, /checkpoint, /restore
 
 ### Artifacts
 
@@ -224,19 +224,25 @@ ${HOOKS_TEXT}
 
 ### Working practices
 
-**Honesty:** If you can't find a doc, file, or function, say "I did not find X. What should I do?" Don't fabricate. If 2-3 attempts fail, escalate with what you tried.
+**Honesty:** If you can't find something, say "I did not find X. What should I do?" Don't fabricate. If 2-3 attempts fail, escalate with what you tried.
 
-**Parallelism:** Move fast. Spin up parallel agents for independent tasks. Run independent tool calls in the same message. Parallelize wherever possible — this is the primary speed lever.
+**Simplicity:** Minimum code that solves the problem. If 200 lines could be 50, rewrite. No abstractions for single-use code. No features beyond what was asked.
 
-**Scope:** Stay focused. If you notice something worth fixing outside scope, mention it at the end, don't silently change it.
+**Surgical changes:** Touch only what's relevant. Don't improve adjacent code, comments, or formatting. Every changed line should trace to the request. Match existing style.
 
-**Completion:** "Done" means verified. Run the tests. Prefer "I ran it and it works" over "should work."
+**Goal-driven:** Define success criteria before coding. Write the test first, then make it pass. For multi-step tasks, state a brief plan with verification for each step.
+
+**Parallelism:** Spin up parallel agents for independent tasks. Run independent tool calls in the same message. This is the primary speed lever.
+
+**Completion:** "Done" means verified. Run the tests. Run the build. Prefer "I ran it and it works" over "should work."
 
 **Failures:** When you fail, say so. Escalate: what you tried, why each failed, what you need.
 
 **User context:** If something looks wrong, ask why before fixing. Users have constraints you may not see.
 
-**Session:** Long conversation? Suggest /compact or a new session. Past 90 min, suggest a break.
+**Planning:** Use plan mode for complex tasks. Break work small enough to complete in under 50% context.
+
+**Session:** /compact at ~50% context. Past 90 min, take a break. Run /checkpoint before ending a session. Run /restore at the start of the next one.
 
 ### Launching sessions
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "uv-suite",
-  "version": "0.12.0",
+  "version": "0.14.0",
   "description": "Portable framework for AI-assisted software development. 10 agents, 9 skills, 5 hooks, 4 personas. Works with Claude Code, Cursor, and Codex.",
   "author": "Utsav Anand",
   "license": "MIT",

package/skills/architect/SKILL.md

@@ -36,3 +36,7 @@ $ARGUMENTS
 !`ls uv-out/specs/*.md 2>/dev/null | head -5 || echo "No specs found"`
 
 !`cat $(ls -t uv-out/specs/*.md 2>/dev/null | head -1) 2>/dev/null | head -80 || echo ""`
+
+### Session checkpoint
+
+!`cat uv-out/checkpoints/latest.md 2>/dev/null | head -40 || echo "No checkpoint"`

package/skills/checkpoint/SKILL.md

@@ -0,0 +1,80 @@
+---
+name: checkpoint
+description: >
+  Save a checkpoint of the current session — what was done, key decisions, current state,
+  and what's next. Use before ending a session, before /compact, or at any natural breakpoint.
+  The next session auto-loads the latest checkpoint.
+argument-hint: "[optional-label]"
+user-invocable: true
+allowed-tools:
+  - Write(*)
+  - Read(*)
+  - Bash(git status *)
+  - Bash(git diff *)
+  - Bash(git log *)
+  - Bash(date *)
+  - Bash(ls *)
+---
+
+## Write a checkpoint to uv-out/checkpoints/
+
+Create the directory `uv-out/checkpoints/` if it doesn't exist.
+
+Write a file named `uv-out/checkpoints/YYYY-MM-DD-HHMM.md` (using the current timestamp).
+
+Also write/overwrite `uv-out/checkpoints/latest.md` with the same content (so the next session can always find the most recent checkpoint).
+
+## Label
+
+$ARGUMENTS
+
+If a label was provided, include it in the filename: `uv-out/checkpoints/YYYY-MM-DD-HHMM-[label].md`
+
+## What to capture
+
+Review the full conversation so far and write a structured checkpoint with these exact sections:
+
+```markdown
+# Checkpoint: [date] [time] [label if provided]
+
+## What was accomplished
+- [Bullet list of concrete things done this session]
+- [Be specific: "Added webhook retry logic to PaymentService" not "worked on payments"]
+
+## Key decisions made
+- [Decision]: [Why] — [What was considered and rejected]
+- [Only include decisions that affect future work]
+
+## Current state
+- Branch: [current git branch]
+- Uncommitted changes: [yes/no, summary if yes]
+- Tests: [passing/failing/not run]
+- Blockers: [any unresolved issues]
+
+## Files modified
+- [List key files changed, not every file]
+
+## What's next
+- [Immediate next step — what the next session should start with]
+- [Remaining tasks from the current Act/plan]
+
+## Context the next session needs
+- [Anything non-obvious that would be lost without this checkpoint]
+- [Workarounds in place, temporary decisions, "this looks wrong but it's intentional because..."]
+- [Environment setup notes if relevant]
+```
+
+## Git state to capture
+
+!`git branch --show-current 2>/dev/null || echo "not a git repo"`
+
+!`git status --short 2>/dev/null | head -20 || echo "no git"`
+
+!`git log --oneline -5 2>/dev/null || echo "no git history"`
+
+## Rules
+
+- Be specific. "Worked on auth" is useless. "Added JWT refresh token rotation with 7-day expiry" is useful.
+- Capture WHY decisions were made, not just what. The next session needs the rationale.
+- Keep it under 80 lines. This isn't a novel — it's a handoff.
+- Every checkpoint overwrites `latest.md` so the next session always finds the freshest state.

package/skills/restore/SKILL.md

@@ -0,0 +1,22 @@
+---
+name: restore
+description: >
+  Restore the latest checkpoint from a previous session. Shows what was done,
+  key decisions, current state, and what's next. Use at the start of a new session.
+user-invocable: true
+allowed-tools:
+  - Read(*)
+  - Bash(ls *)
+---
+
+## Latest checkpoint
+
+!`cat uv-out/checkpoints/latest.md 2>/dev/null || echo "No checkpoint found. Run /checkpoint to create one."`
+
+## All checkpoints
+
+!`ls -la uv-out/checkpoints/*.md 2>/dev/null | tail -10 || echo "No checkpoints directory"`
+
+## Instructions
+
+Read the checkpoint above. Summarize it to the user in 3-4 sentences: what was done, what's the current state, and what's next. Then ask: "Ready to pick up from here, or do you want to take a different direction?"

package/skills/review/SKILL.md

@@ -51,3 +51,7 @@ $ARGUMENTS
 ### Acts plan
 
 !`cat uv-out/architecture/acts-plan.md 2>/dev/null | head -60 || echo "No acts plan found"`
+
+### Session checkpoint (what's in progress)
+
+!`cat uv-out/checkpoints/latest.md 2>/dev/null | head -40 || echo "No checkpoint"`

package/skills/write-tests/SKILL.md

@@ -48,3 +48,7 @@ $ARGUMENTS
 ### Acts plan (current task context)
 
 !`cat uv-out/architecture/acts-plan.md 2>/dev/null | head -40 || echo "No acts plan found"`
+
+### Session checkpoint
+
+!`cat uv-out/checkpoints/latest.md 2>/dev/null | head -40 || echo "No checkpoint"`