spinosa-overseer

Audits research workspace coverage and generates gap analysis reports from workspace sources.

7|Updated May 13, 2026
One-click install
npx skills add https://github.com/medialab/spinosa --skill spinosa-overseer-medialab
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: spinosa-overseer
Source: https://github.com/medialab/spinosa/tree/main/workspace-template/.opencode/skills/spinosa-overseer
Command: npx skills add https://github.com/medialab/spinosa --skill spinosa-overseer-medialab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Research workspaces accumulate maps, raw documents, and agent reports over time, making it hard to know which topics, files, and agents have been neglected. This Skill performs retrospective coverage audits to surface gaps, stale areas, and underutilized agents. ## Core Features & Use Cases - Coverage Auditing: Matches dictionary concepts against orchestrator notes and session history to classify topics as covered, stale, or not covered. - Gap Detection: Identifies untouched maps, unvisited raw directories, outdated reports, and underutilized agents, with optional external session-log forensics. - Actionable Recommendations: Produces adversarial multi-agent probe proposals, re-mapping suggestions, and orchestrator advisories written to agent_reports/. - Use Case: After weeks of interview transcript analysis, run a coverage audit to discover that two raw directories were never visited and three dictionary concepts have zero sessions, then receive prioritized dispatch recommendations. ## Quick Start Ask the agent to run a coverage audit of the workspace and write a gap report to agent_reports after the workspace has started.

Frequently Asked Questions about spinosa-overseer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a coverage audit on a research workspace?▼

Invoke the overseer after the workspace reaches workspace_started status, either through the orchestrator's periodic counter or an explicit user request. It reads the dictionary, orchestrator notes, maps, and reports, then writes a coverage report to agent_reports/.

What does a coverage gap report include?▼

The report includes a coverage gauge, topic and corpus coverage matrices, agent utilization stats, identified concept and process gaps, and adversarial recommendations with suggested agent dispatches. It is written as a Markdown file in agent_reports/.

Can the coverage audit run during startup indexing?▼

No. The agent refuses to run while setup_status is cli_started and returns overseer_skipped. Coverage audits are only permitted after the workspace has fully started to avoid interfering with indexing.

Does the coverage audit require external session logs?▼

No. External Codex or OpenCode session logs are optional enrichment used only when explicitly requested. If logs are unavailable, the audit soft-fails to in-workspace-only analysis and notes the limitation in the report.

What are the limitations of file-level coverage analysis?▼

File-level percentages such as thin coverage below 20% require external session-log data. Without logs, the audit reports directory-level proxies from notes and inventory only, and marks file-level depth as unavailable.