user-feedback

To route user feedback automatically, this Skill parses natural language input, classifies it into categories like project convention or agent behavior, and updates MEMORY.md, SKILL.md files, and agent prompts directly. It reads target files to avoid duplicates and replaces contradictory rules.

Updating repository conventions from feedback safely involves reading existing target files to check for duplicates before appending or modifying. The Skill updates rules in place within MEMORY.md or SKILL.md files and produces a summary of applied changes for reviewer approval.

Propagating workflow preference changes requires identifying affected skills and agents, then updating their instructions or prompts as needed. This Skill classifies the preference, applies it to relevant project files, and generates a concise summary of all modifications for approval.

Yes, you can apply DFM and PCB manufacturing rules to project memory using natural language parsing. The Skill maps your input into the DFM/PCB rule category, reads target files to prevent duplication, and appends or updates the new guidance in MEMORY.md.

Yes, this feedback routing approach works for updating agent behavior during development. It classifies feedback specifically as agent behavior changes, reads the existing agent definitions to avoid conflicts, and updates the prompts accordingly before outputting a change summary.