feedback-protocol

install
source · Clone the upstream repo
git clone https://github.com/obsfx/cc-feedback
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/obsfx/cc-feedback "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/feedback-protocol" ~/.claude/skills/obsfx-cc-feedback-feedback-protocol && rm -rf "$T"
manifest: skills/feedback-protocol/SKILL.md
source content

Feedback Protocol

@feedback
annotations are inline source code comments that describe desired changes. They use the file's native comment style (
//
,
#
,
--
,
/* */
,
<!-- -->
).

Format

// @feedback: <desired change>
// @agent-response: <what was done>

The

@agent-response
line is added directly below the
@feedback
line using the same comment style after implementing the change.

Status

  • Unaddressed:
    @feedback
    with no
    @agent-response
    below it
  • Addressed:
    @feedback
    followed by
    @agent-response
    on the next line
  • Only unaddressed feedbacks need action

Workflow

  1. Discover: Grep for
    @feedback
    in source files, exclude non-source dirs
  2. Parse: Extract file path, line number, feedback text, comment style
  3. Context: Read surrounding code to understand what the feedback refers to
  4. Implement: Make the requested change
  5. Respond: Add
    @agent-response: <summary>
    — max 300 characters. If unable:
    @agent-response: skipped - <reason>

Proactive Behavior

When encountering

@feedback
while working on other tasks: note them to the user and offer to address if related. Do not address silently.

Reference

For comment style details and edge cases (multi-line blocks, inline placement, nested comments): see

references/comment-patterns.md
.