Run any Skill in Manus with one click

session-digest

Stars28

Forks1

UpdatedMarch 7, 2026 at 07:15

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

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

hua-bang

hua-bang/pulse-agent

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Data ScientistsComputer and Mathematical Occupations·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"}

More from this repository

same repository

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"}