Run any Skill in Manus with one click

session-summarizer

Stars2

Forks0

UpdatedJune 3, 2026 at 22:10

Leaving-agent protocol. Produces a compressed, decision-first handoff that preserves intent and next steps while stripping execution noise.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

pbuckles22

pbuckles22/Rust_Julia_FFI_Flutter_Template

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Software DevelopersComputer and Mathematical Occupations·SOC 15-1252

SKILL.md

readonly

name	session-summarizer
description	Leaving-agent protocol. Produces a compressed, decision-first handoff that preserves intent and next steps while stripping execution noise.

Session Summarizer — Leaving Agent Protocol

Use this skill when ending a session, reducing context, handing work to a new agent, or preparing a "feature-agent" to pause safely.

Goal: transfer working state (decisions, rationale, next steps) with minimal tokens.

Progressive summarization (what to keep vs strip)

Always keep (highest value per token):

Decisions (what we chose)
Rationale (why we chose it)
Next steps (what to do next, in order)
Acceptance criteria / validation (how to know it's done)

Keep only summary-level:

Actions taken (headline only)
File paths changed (key files only)

Strip unless directly decision-relevant:

Long logs
Step-by-step execution transcripts
Repeated context already captured in tracked docs

Handoff note budget

Default target: ≤ 500 words (≈ 700 tokens).

If you exceed the budget, remove execution details first.

Handoff note structure (required)

Write the handoff note using this structure (mirrors the repo template):

Filename rules (mandatory)

Prefer: .cursor/handoff/NNNN-handoff-YYYY-MM-DD_HHmm.md
Optional second location: doc/handoff/NNNN-HANDOFF-YYYY-MM-DD_HHmm.md
NNNN must be new and monotonic (0001, 0002, …). Never reuse a number.
Never edit a prior handoff file to "update" it—write a new one.

Note contents

TL;DR (1–2 sentences): current state + the one key thing the next agent must know
Decisions made: bullet list (decision + rationale)
Done this session: summary bullets (no logs)
Next steps (prioritized): 1–5 steps; each step should be verifiable
Blockers / open questions: only items that prevent safe progress
Durable docs updated: list tracked docs updated (or "none")
Key files: only the handful that matter next

"Green and Clean" exit check

Before ending:

Are acceptance criteria satisfied (or clearly not yet)?
Were the right validation tiers run (Tier 1 minimum where defined)?
Are durable decisions written into tracked docs (not only in the note)?
Is the note short enough to prevent context bloat?

More from this repository

same repository

code-quality-gate

pbuckles22/Rust_Julia_FFI_Flutter_Template

Change-aware quality pass: diff scope, readability, function size and control flow, comments where non-obvious, and alignment with project linting. Use before merge or with handoff when the user wants industry-style maintainability beyond what tests assert.

2026-06-032

code-reviewer

pbuckles22/Rust_Julia_FFI_Flutter_Template

Review code changes for correctness, security, project conventions, and maintainability. Use when reviewing a diff, before commit, or when the user asks for a code review.

2026-06-032

context-bootstrapper

pbuckles22/Rust_Julia_FFI_Flutter_Template

Receiving-agent protocol. Boots a new agent into the minimum correct context (select + isolate) and produces a clear next-step plan without guessing.

2026-06-032

eval-engineer

pbuckles22/Rust_Julia_FFI_Flutter_Template

Define and maintain lightweight evaluations (acceptance criteria, regression checks, golden tests) so "green" is objective and repeatable.

2026-06-032

game-readiness

pbuckles22/Rust_Julia_FFI_Flutter_Template

Validates a game or interactive product for playability, clarity, feedback, and scale-readiness. Use at milestones, before large scope changes, or when the user asks if the product feels ready or is ready to scale.

2026-06-032

github-feature-workflow

pbuckles22/Rust_Julia_FFI_Flutter_Template

Short-lived feature branches; TDD + lint + merge-ready command as exit criteria before commit; push and merge to main (or user-directed flow). Do not default to asking the user to open a PR. Use when implementing a feature or non-trivial fix, when the user asks for branch/git workflow, or after substantial edits that should not stay uncommitted.

2026-06-032

name	session-summarizer
description	Leaving-agent protocol. Produces a compressed, decision-first handoff that preserves intent and next steps while stripping execution noise.

Session Summarizer — Leaving Agent Protocol

Use this skill when ending a session, reducing context, handing work to a new agent, or preparing a "feature-agent" to pause safely.

Goal: transfer working state (decisions, rationale, next steps) with minimal tokens.

Progressive summarization (what to keep vs strip)

Always keep (highest value per token):

Decisions (what we chose)
Rationale (why we chose it)
Next steps (what to do next, in order)
Acceptance criteria / validation (how to know it's done)

Keep only summary-level:

Actions taken (headline only)
File paths changed (key files only)

Strip unless directly decision-relevant:

Long logs
Step-by-step execution transcripts
Repeated context already captured in tracked docs

Handoff note budget

Default target: ≤ 500 words (≈ 700 tokens).

If you exceed the budget, remove execution details first.

Handoff note structure (required)

Write the handoff note using this structure (mirrors the repo template):

Filename rules (mandatory)

Prefer: .cursor/handoff/NNNN-handoff-YYYY-MM-DD_HHmm.md
Optional second location: doc/handoff/NNNN-HANDOFF-YYYY-MM-DD_HHmm.md
NNNN must be new and monotonic (0001, 0002, …). Never reuse a number.
Never edit a prior handoff file to "update" it—write a new one.

Note contents

TL;DR (1–2 sentences): current state + the one key thing the next agent must know
Decisions made: bullet list (decision + rationale)
Done this session: summary bullets (no logs)
Next steps (prioritized): 1–5 steps; each step should be verifiable
Blockers / open questions: only items that prevent safe progress
Durable docs updated: list tracked docs updated (or "none")
Key files: only the handful that matter next

"Green and Clean" exit check

Before ending:

Are acceptance criteria satisfied (or clearly not yet)?
Were the right validation tiers run (Tier 1 minimum where defined)?
Are durable decisions written into tracked docs (not only in the note)?
Is the note short enough to prevent context bloat?