From dev
Apply tagged review comments (e.g. [MK], [REV]) in the current document immediately, using per-tag interpretation rules from `~/.mstack/dev/feedback-tags.json`. Use when the user says "apply feedback", "apply MK comments", "apply review notes", "apply my edits", or wants tagged comments incorporated directly into a document without an approval gate.
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev:apply-feedbackThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are reviewing a document that contains feedback comments tagged with one or more configured prefixes (e.g., `[MK]`, `[REV]`). Tags and their interpretation rules are configured per-user in `~/.mstack/dev/feedback-tags.json`.
You are reviewing a document that contains feedback comments tagged with one or more configured prefixes (e.g., [MK], [REV]). Tags and their interpretation rules are configured per-user in ~/.mstack/dev/feedback-tags.json.
Run this bash to check for the config file:
mkdir -p ~/.mstack/dev
CONFIG=~/.mstack/dev/feedback-tags.json
if [ -f "$CONFIG" ] && [ -s "$CONFIG" ]; then
cat "$CONFIG"
else
echo "MISSING"
fi
MISSING (first run)Create the file with the default MK entry. Use the Write tool to write ~/.mstack/dev/feedback-tags.json with this content:
{
"tags": [
{
"tag": "MK",
"from": "Document author (initials = MK for Mayank)",
"content": "Direct edit instructions, factual corrections, rephrasing requests, structural changes",
"action": "Apply edits as written; treat as imperative; only push back if ambiguous"
}
]
}
Tell the user verbatim:
Saved your feedback tag config to
~/.mstack/dev/feedback-tags.json. Pre-seeded withMKas the default. Edit this file directly to add, modify, or delete tags later.
Use AskUserQuestion: "Want to add another tag now? You can also add more later by editing the file." with options:
If the user picks B, ask 4 questions in sequence (one AskUserQuestion call each, free-text answers):
Read the current config, append the new tag object to the tags array, write the updated file. Then ask "Add another?" again — loop until user says no.
Proceed to Step 1 once setup completes.
Parse it. Build a list of all configured tags from the tags array. Each tag has fields tag, from, content, action.
If JSON is malformed, warn the user (e.g., "feedback-tags.json is invalid JSON, falling back to default MK tag for this run"), then use the default MK tag with the same defaults as the first-run pre-seed.
For each tag from the loaded config, search the document case-insensitively for the bracketed pattern. For example, if tags MK and REV are configured, search for [MK], [mk], [Mk], [REV], [rev], [Rev], etc.
Use the Grep tool with the -i flag, or grep -i "\[mk\]" style commands. Aggregate matches across all configured tags.
If no matches found for any tag, tell the user: "No comments found for any configured tag: TAG1, TAG2, …" listing the configured tag prefixes, and stop.
For each match, apply the requested change without asking permission, using the per-tag action and content fields to inform interpretation:
action says "Apply edits as written; treat as imperative" → just apply the edit.action says "Treat as suggestions; flag for discussion" → in this apply-feedback skill, still apply (this skill's posture is apply-now), but be more conservative; if the suggestion is open-ended ("consider rewording"), make the most reasonable edit and note it in your final summary.Apply edits by:
Ensure changes integrate smoothly with existing content. Preserve overall structure and tone. Match the document's voice when adding content.
action and content fields as guidance for how aggressively/literally to applyWith MK configured (action: "Apply edits as written"):
Before:
The system will use the brand_mentions table [MK] Rename to response_metrics.
After:
The system will use the response_metrics table
Apply all configured-tag feedback now and update the document.
npx claudepluginhub mayank-io/mstack --plugin devGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.