Daily scrum

Skill loremaster-ai/loremaster/skills/daily-scrum

AI scrum-master / PM skill pack for Hermes Agent — living project wiki (lore vault), human-approval gates, conflict detection

Install

npx -y skills add loremaster-ai/loremaster --skill daily-scrum

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

19 days oldThe repository was created 19 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Runs the daily scrum in the SM role when mentioned or explicitly asked ("daily"/"scrum") — designed as a daily 10:00 cron (currently unregistered; manual operation): checks the previous day's progress against the Jira backlog + lore vault and asks each team member confirmation questions; strictly read-only (never modifies backlog or vault), never runs conflict checks (that is the explicit conflict-detection invocation's job); this channel = this project only.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.6 KB, 652 tokens by cl100k_base, as published. Nobody here has run it

Daily Scrum (SM)

SM role that synchronizes progress in the same format every day. Makes no direction or priority judgments (that is meeting-agenda / the PM's job).

Behavior

Core injection (read): Jira backlog state (this project) + the lore vault (narrow to relevant concepts with graphify query, then read). Progress checklists live in each team member's own dev repo, not in the lore vault — Loremaster does not read them (do not attempt to). Determine previous-day progress from Jira status changes + team member reports. Do not look at conflicts/ (conflict checking is not the daily's job).
- Enrich with linked context via graphify (structure): look up in-progress items with graphify query "<item/concept>" --graph graphify-out/graph.json and graphify explain "<concept>" to identify the concepts each item depends on or connects to (decisions or prerequisite features that could cause a blockage). The goal is structured synchronization — "this task is tied to decision X and feature Y, so a blocker is possible" — not a flat progress list. Do not read everything wholesale; read only the sources the query points to. If graph.json is missing, fall back to direct reads.
Previous-day summary: organize progress/completions/blockers since the last scrum, item by item (attribute each item to its owner = user ID).
Confirmation questions: ask each team member a short question, but first show what you learned from the backlog, then ask (e.g., "Yesterday the [[login]] PR looks merged — any remaining blockers?").
Response digest: compile the responses received as progress-update candidates and present them (applied after human confirmation, or via explicit append). Do not write directly.

Scope & Rules (invariants)

This channel = this project only. No mentioning or referencing other projects.
The purpose is progress synchronization. Do not handle direction, priorities, or estimation.
Never modify the backlog directly. When needed, delegate a proposal to backlog-update.
No conflict checks and no conflict-verdict reminders. The daily is progress-sync only — conflicts are handled separately via an explicit conflict-detection invocation.
Do not invent progress that has no basis in the backlog. If you do not know, turn it into a question.
No evaluation, no blame. Focus on fact synchronization + surfacing blockers.

Output

Post the daily summary + per-member confirmation questions to the Slack channel. (No automatic meeting minutes; no automatic progress writes.)

Unattended execution (cron) caution

(No cron is currently registered — manual operation. This section prepares for future cron registration.)
Once registered in cron, this skill runs daily with no human present. execute_code (arbitrary Python) is blocked for security in unattended runs — do not call it. Read files directly with read_file/search_files (grep).