Execute qualquer Skill no Manus
com um clique

Execute qualquer Skill no Manus com um clique

session-digest

Estrelas28

Forks1

Atualizado7 de março de 2026 às 07:15

Summarize sessions across a multi-day range by paging session_summary in chunks, output bilingual, and write back stable soul traits.

Instalação

Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.

Executar no Manus

Fonte

hua-bang

hua-bang/pulse-agent

Abrir repositório GitHub Ver repositórios do creator

Download

Executar no Manus

Ocupações relacionadasSOC

Baseado na classificação ocupacional SOC

Cientistas de dadosInformática e Matemática·SOC 15-2051

SKILL.md

readonly

name	session-digest
description	Summarize sessions across a multi-day range by paging session_summary in chunks, output bilingual, and write back stable soul traits.
description_zh	通过分段调用 session_summary 汇总多天会话，输出中英双语摘要，并在稳定特征出现时写回 soul。
version	1.0.0
author	Pulse Coder Team

Session Digest Skill

Summarize sessions over a long date range without blowing context by paging session_summary in windows and aggregating the results.

When to Use

You need summaries for many days (e.g., 14/30/90 days)
Single tool output is too large to fit in context
You want a high-level digest with minimal raw excerpts

Inputs (JSON)

{
  "days": 30,
  "chunkDays": 7,
  "scope": "owner",
  "includeUserMessages": true,
  "includeAssistantMessages": true,
  "maxMessagesPerSession": 120,
  "focus": "optional: emphasize PRs, configs, incidents, or product decisions",
  "format": "themes"
}

days (default 7): total days to summarize (UTC, counting back from today).
chunkDays (default 7, max 30): size of each window.
scope: owner (cross-channel) or platform (current channel).
maxMessagesPerSession: reduce to avoid overlong excerpts.
format: themes or timeline.

Required Execution Flow

Compute totalDays from input (default 7).
Compute chunkDays (default 7, clamp 1..30).

For each window index i = 0..:

offsetDays = i * chunkDays
windowDays = min(chunkDays, totalDays - offsetDays)

Call session_summary with:

{
  "days": windowDays,
  "offsetDays": offsetDays,
  "scope": "owner",
  "includeUserMessages": true,
  "includeAssistantMessages": true,
  "maxMessagesPerSession": 120
}

Summarize this window to 3-6 bullets, no raw excerpt dumps.

Aggregate across windows:
- Deduplicate repeated items.
- Highlight: PRs/MRs, code changes, config changes, incidents, decisions.
Produce bilingual output (EN + ZH) with the format below.
Soul writeback (no read):
- Do NOT call memory_recall.
- Infer 1-2 stable traits only when the same tendency appears in >= 2 sessions or windows.
- Record with memory_record using kind: "soul" and a short, neutral sentence (no sensitive or user-facing text).
- Do not mention the soul writeback in the visible output.

Output Format

Start with the date range (UTC).
Provide English then Chinese sections.
Keep each language concise (target <= 1200 chars; avoid long excerpts).

Example:

Date Range (UTC): 2025-03-01 to 2025-03-07

EN
Overall:
- ...
- ...
Per-window:
- 2025-03-07: ...
- 2025-03-06: ...
Notable:
- ...

ZH
总体:
- ...
- ...
分段:
- 2025-03-07: ...
- 2025-03-06: ...
要点:
- ...

Output Rules

Never paste the raw session_summary.sessions[].excerpt list.
Prefer concise phrasing and avoid low-signal chat.
If a window has no sessions, say so explicitly.
Keep total output under ~300 lines.

Example Invocation

/skills session-digest {"days": 30, "chunkDays": 7, "scope": "owner"}

Mais deste repositório

mesmo repositório

validate-canvas-change

hua-bang/pulse-agent

Select and run proportionate validation for changes to apps/canvas-workspace. Use when iterating on Canvas Workspace code, finishing a feature or fix, preparing release evidence, or deciding whether a full performance report is necessary.

2026-07-2628

canvas-bootstrap

hua-bang/pulse-agent

Deep-research a topic and build a structured Pulse Canvas workspace with approved research depth, source-backed findings, progressive or final canvas creation, spatially organized frames, content nodes, and connections. Use when the user asks to bootstrap, generate, research, organize, or build an AI-created canvas.

2026-07-2428

canvas-frame-research

hua-bang/pulse-agent

Extend an existing Pulse Canvas frame with local, source-backed research while preserving the surrounding canvas. Use when the user asks to research, enrich, expand, verify, update, or add sources/details inside a selected or named frame, rather than bootstrapping a whole new canvas.

2026-07-2428

canvas

hua-bang/pulse-agent

Operate Pulse Canvas workspaces — read user-curated context, write results, create nodes

2026-07-1928

add-canvas-capability

hua-bang/pulse-agent

Add or change a Pulse Canvas capability that must be available through structured agent tools and the pulse-canvas CLI. Use for cross-workspace Capability Runtime changes spanning apps/canvas-workspace, packages/canvas-cli, or packages/cli, including new live read, operate, or unsafe application actions.

2026-07-1928

perf-report

hua-bang/pulse-agent

Run the canvas-workspace performance evaluation, publish the latest static dashboard through the local nginx/Cloudflare Tunnel route, capture a screenshot, and summarize the report. Use when the user asks to run Pulse Canvas/canvas-workspace perf checks, deploy the performance dashboard, refresh https://jasperhu.art/apps/canvas-perf/, or send a dashboard screenshot from Feishu/remote-server.

2026-07-1228

Session Digest Skill

Summarize sessions over a long date range without blowing context by paging session_summary in windows and aggregating the results.

When to Use

You need summaries for many days (e.g., 14/30/90 days)

Single tool output is too large to fit in context

You want a high-level digest with minimal raw excerpts

Inputs (JSON)

{ "days": 30, "chunkDays": 7, "scope": "owner", "includeUserMessages": true, "includeAssistantMessages": true, "maxMessagesPerSession": 120, "focus": "optional: emphasize PRs, configs, incidents, or product decisions", "format": "themes" }

days (default 7): total days to summarize (UTC, counting back from today).

chunkDays (default 7, max 30): size of each window.

scope: owner (cross-channel) or platform (current channel).

maxMessagesPerSession: reduce to avoid overlong excerpts.

format: themes or timeline.

Required Execution Flow

Compute totalDays from input (default 7).

Compute chunkDays (default 7, clamp 1..30).

For each window index i = 0..:

offsetDays = i * chunkDays
windowDays = min(chunkDays, totalDays - offsetDays)

Call session_summary with:

{
  "days": windowDays,
  "offsetDays": offsetDays,
  "scope": "owner",
  "includeUserMessages": true,
  "includeAssistantMessages": true,
  "maxMessagesPerSession": 120
}

Summarize this window to 3-6 bullets, no raw excerpt dumps.

Aggregate across windows:

Deduplicate repeated items.
Highlight: PRs/MRs, code changes, config changes, incidents, decisions.

Produce bilingual output (EN + ZH) with the format below.

Soul writeback (no read):

Do NOT call memory_recall.
Infer 1-2 stable traits only when the same tendency appears in >= 2 sessions or windows.
Record with memory_record using kind: "soul" and a short, neutral sentence (no sensitive or user-facing text).
Do not mention the soul writeback in the visible output.

Output Format

Start with the date range (UTC).

Provide English then Chinese sections.

Keep each language concise (target <= 1200 chars; avoid long excerpts).

Example:

Date Range (UTC): 2025-03-01 to 2025-03-07 EN Overall: - ... - ... Per-window: - 2025-03-07: ... - 2025-03-06: ... Notable: - ... ZH 总体: - ... - ... 分段: - 2025-03-07: ... - 2025-03-06: ... 要点: - ...

Output Rules

Never paste the raw session_summary.sessions[].excerpt list.

Prefer concise phrasing and avoid low-signal chat.

If a window has no sessions, say so explicitly.

Keep total output under ~300 lines.

Example Invocation

/skills session-digest {"days": 30, "chunkDays": 7, "scope": "owner"}