Create read-only research sessions and output ANALYSIS-FILE.md and CONCLUSIONS.md.
npx skills add https://github.com/Tacuchi/agent-workflow-cli --skill research-tacuchi
Please help me install this Agent Skill. Skill: research Source: https://github.com/Tacuchi/agent-workflow-cli/tree/main/skills/w/roles/research Command: npx skills add https://github.com/Tacuchi/agent-workflow-cli --skill research-tacuchi
Provides on-demand investigation sessions that read-only inspect the workspace and repositories to gather evidence, then produce ANALYSIS-FILE and CONCLUSIONS, preventing production artifacts from being altered.
Create a session research for a given question and let the system generate ANALYSIS-FILE.md and CONCLUSIONS.md from the gathered evidence.
High-intent search queries and answers about installing and using this skill.
On-demand research sessions inspect the workspace using read-only operations like Read, Grep, Glob, and MCP queries to gather evidence. This read-only approach ensures production artifacts remain completely unaltered during the investigation process.
To generate structured conclusions, the research session writes raw gathered evidence to ANALYSIS-FILE.md and synthesizes findings into CONCLUSIONS.md. Sessions close as either closed or inconclusive based on the verifiable evidence collected from the workspace context.
Automated research sessions apply across spec-refine-loop, plan-new-loop, plan-exec-loop, and quick-loop workflows. The system reads code and MCP data to decide whether to investigate autonomously or prompt the human, grounding workflow decisions with verifiable evidence.
The best way to automate evidence collection is creating an on-demand research session targeting a specific question. The system applies read-only operations to inspect code and MCP data, exporting queries only when needed to preserve production artifacts.
Use an automated investigation session when you need to ground workflow decisions with verifiable evidence across AI loops. It enforces safety rules and prevents accidental modifications, making it safer than manual codebase analysis for producing structured conclusions.
lucaswallInvestigate Railway logs and codebase to report findings on API, deployment, and configuration issues.
lee-toOrchestrate iterative artifact refinement through PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE phases.
mattbutlerengineeringCompare installed tools against WORKFLOW.md to report dev loop gaps.
lemigAnalyze a web application with report-only tests and produce a structured bug report.
