Agent Skills by QS
Showing 12 vetted skills indexed across 1 GitHub repositories.
QSchlegelqfeed
Files gated findings as tasks into a team queue via the session-viz tasks API.
QSchlegelqsetup
Connects a machine to a session-viz workspace via browser-based OAuth sign-in.
QSchlegelqteam
Resolves Obsidian wikilinks across federated vaults and manages task handoffs via the session-viz MCP server.
QSchlegelqruns
Audits scheduled runs and subagents from local Claude Code, Codex, and Cursor transcripts.
QSchlegelqpact
Analyzes the current Claude Code session and renders an interactive HTML report with a tuned /compact instruction.
QSchlegelqdoctor
Audits a repository's Claude Code configuration against the user's other repos.
QSchlegelqship
Detects repeated prompts across sessions and generates slash command files for recurring rituals.
QSchlegelqtrends
Analyzes Claude Code session transcripts for friction trends and renders an interactive HTML report.
QSchlegelqshare
Share selected projects, sessions, or runs with a workspace after privacy review.
QSchlegelqbl
Files and retrieves project backlog items with explicit ordering criteria and dependency tracking.
QSchlegelqcontrib
Sends a local delivery ledger as nine bounded, privacy-redacted columns to a shared reference endpoint.
QSchlegelqcost
Analyzes token spend across agent runs, separating cache-read from generated output costs.
Frequently Asked Questions About QS
FAQPage SchemaWhat tasks can I perform with QS's skills?▼
You can analyse Claude Code sessions for prompting quality and friction, track token spend by cache-read versus generated output, inspect a delivery ledger of scheduled runs and subagents, convert repeated prompts into slash commands, manage a bidirectional backlog, and hand tasks between teammates via shared Obsidian vaults.
Who are these skills designed for?▼
Individual engineers and teams running Claude Code locally who want visibility into session quality, token costs, and autonomous run outcomes. Team leads benefit from qteam, qshare, and qfeed for federated vaults, controlled sharing, and gated task queues across projects.
How do the skills handle privacy and sharing?▼
Nothing is shared by default. qshare requires explicit opt-in per project, session, or run, and qcontrib transmits only nine bounded columns — task class, ISO week, terminal state, and log2 token buckets — with no prompts, paths, repo names, or fine-grained timestamps.
How does authentication and setup work?▼
qsetup connects a machine to the session-viz workspace through browser sign-in only; no token is ever typed, pasted, or shown in the terminal. Most commands are invoked explicitly via slash commands, while read-only stdout skills like qruns, qdoctor, and qcost are model-invocable.
Which skills are read-only versus write or network operations?▼
qruns, qdoctor, and qcost only read the local machine and print to stdout, so they are model-invocable. qpact and qtrends open browser previews, qship --write creates command files, and the five network commands (qsetup, qteam, qshare, qfeed, qcontrib) require explicit slash invocation.