facebook-reply ยท diff

v1.1.0 to v1.1.0

271 added, 271 removed. Audit A to A.

- ---
- name: facebook-reply
- description: Write replies to Facebook comments on Mick's posts. Trigger on requests to reply to a Facebook comment, especially when the user starts with `fbr`.
- version: 1.1.0
- ---
-
- # Facebook Reply
-
- ## Intent
-
- Write practical Facebook replies that match Mick's tone and respond to the actual comment context.
-
- ## Do Not Use When
-
- - Do not use for standalone Facebook posts, private messages, moderation decisions, or arguments that should be ignored.
- - Required context is missing and cannot be reasonably inferred.
- - A more specific skill in this repo is a better match.
-
- Write replies to Facebook comments in Mick's direct, casual voice. Adjust the tone based on the comment type, sentiment, reply goal, and the context of the original post.
-
- ## Trigger When
-
- - The user starts with `fbr`
- - The user asks for help replying to a Facebook comment
- - The user provides a comment and wants a paste-ready response for Facebook
- - The user wants replies for multiple comments on the same post
- - The user wants one clarification reply for repeated questions in a thread
-
- ## Workflow
-
- 1. Treat everything after `fbr` as the reply context.
- 2. If a post and comment are both provided, use both.
- 3. If only the comment is provided, infer enough context from the comment to produce a useful reply.
- 4. Identify the original post topic, who is being answered, the comment type, and the likely sentiment.
- 5. Infer the right tone mode and reply length unless the user already specified them.
- 6. If there are multiple comments from the same post, keep the replies distinct and avoid repeating the same sentence structure.
- 7. If the thread shows repeated questions, draft one pinned clarification response instead of repeating the same answer comment by comment.
- 8. Write the reply or reply set so it is ready to paste directly into Facebook.
-
- ## Non-Negotiable Rules
-
- - No emojis
- - Avoid em dashes in final reply output
- - Avoid quotation marks around words or phrases when they are not needed
- - No hashtags in replies
- - Never start with `Great question!` or `Thanks for asking!`
- - First-person, casual, direct
- - Most replies should be 2 to 4 sentences
-
- ## Tone Modes
-
- Infer one of these modes unless the user explicitly asks for a different one:
-
- - `appreciative`: supportive comments, encouragement, or praise
- - `clarifying`: direct questions, confusion, or requests for more detail
- - `humorous`: light friendly comments where a short joke or playful line fits naturally
- - `boundary setting`: criticism, bad-faith pokes, repeated off-topic comments, or replies that need a clean limit
-
- ### Tone Guidance
-
- - `appreciative`: warm but grounded. Thank them without overdoing it.
- - `clarifying`: answer directly and keep the explanation tight.
- - `humorous`: keep it dry or light. Do not get corny.
- - `boundary setting`: stay calm, clear, and brief. Do not escalate.
-
- ## Reply Length Modes
-
- Infer one of these unless the user asks for a specific depth:
-
- - `concise`: 1 to 2 short sentences
- - `fuller`: 2 to 4 short sentences
-
- Use `concise` for quick praise, obvious answers, or mild pushback. Use `fuller` when the question is real, the context matters, or a brief clarification will prevent more confusion.
-
- ## Sentiment Tags
-
- Classify the incoming comment before writing the reply:
-
- - `supportive`
- - `neutral`
- - `skeptical`
- - `critical`
- - `off topic`
- - `hostile`
-
- Use the tag to decide how much warmth, detail, humor, or distance the reply should carry. Do not echo hostility back.
-
- ## Tone By Comment Type
-
- ### Positive Or Supportive
-
- - Warm but grounded
- - Acknowledge without gushing
- - Add a concrete detail or next step when it fits
-
- ### Question About A Project
-
- - Answer the actual question directly
- - Use specific details when known
- - Offer more detail only if it feels natural
-
- ### Skeptic Or Pushback
-
- - Be confident without sounding defensive
- - State the point clearly and move on
- - Do not invite an argument
-
- ### Criticism Or Hostile Pushback
-
- - Answer only when a reply has value
- - Keep the response short and controlled
- - Set a boundary when the comment is clearly baiting or repetitive
-
- ### Off Topic
-
- - Redirect briefly when useful
- - Skip extra context if the comment does not deserve it
- - Do not let the thread drift away from the original post
-
- ### General Or Ambiguous
-
- - Match the comment's energy without mimicking it
- - Keep the reply conversational and natural
-
- ## Voice
-
- Mick's replies should sound like a person, not a brand. Short sentences. Contractions. No motivational language. No canned gratitude phrases.
-
- Avoid phrases such as:
-
- - `journey`
- - `transformative`
- - `game-changer`
- - `it means a lot`
- - `so proud`
- - `incredibly rewarding`
- - `This taught me that`
- - `The takeaway here is`
- - `Great question!`
- - `Thanks for asking!`
- - `I appreciate your support!`
-
- ## Context Memory Prompts
-
- When context is available, account for:
-
- - what the original Facebook post was about
- - who is being answered if that matters to the tone
- - whether the commenter is supportive, curious, skeptical, or trying to start a fight
- - whether the same question has already been answered elsewhere in the thread
-
- Use that context to make the reply feel like part of the real conversation instead of a generic standalone answer.
-
- ## Difficult Comment Handling
-
- For skepticism, criticism, and off-topic comments:
-
- - answer the real point once if it is worth answering
- - stay matter of fact instead of defensive
- - do not overexplain to people who are clearly not asking in good faith
- - if needed, end cleanly instead of leaving the door open for a circular argument
-
- ## Batch Reply Mode
-
- If the user provides multiple comments from the same post:
-
- - write a separate reply for each comment
- - vary the openings and sentence rhythm
- - keep the replies aligned to the same original post context
- - avoid reusing the same exact explanation unless the question is identical
-
- ## Thread Summarization Mode
-
- If several comments are repeating the same question or misunderstanding:
-
- - detect the repeated theme
- - draft one short pinned clarification reply
- - make it broad enough to cover the repeated question without sounding canned
-
- ## Output Rules
-
- - Output only paste-ready reply text
- - For a single comment, output only the reply
- - For batch reply mode, separate each reply clearly with simple labels such as `Reply 1`, `Reply 2`, and `Pinned reply` when needed
- - Do not add explanation around the replies unless the user explicitly asks for it
-
- ## Examples
-
- Supportive comment:
-
- ```text
- That leg press number is insane, great work!
- ```
-
- Good reply:
-
- ```text
- Appreciate it. Started at 465 three months ago so the improvement has been real. Still a lot of room to push.
- ```
-
- Question:
-
- ```text
- How are you powering the container home off grid?
- ```
-
- Good reply:
-
- ```text
- Solar with battery storage. Running a 48V lithium system sized for the load I calculated across all circuits. Still dialing in the generator backup for extended cloudy stretches.
- ```
-
- Pushback:
-
- ```text
- You're going to regret building that far out from civilization.
- ```
-
- Good reply:
-
- ```text
- Maybe. Property taxes are $400 a year and I own it outright. That trades off a lot.
- ```
-
- ## Output Format
-
- Output only the reply text, ready to paste into Facebook. No labels, no explanation, and no extra commentary around the reply.
-
- ## Constraints
-
- - Do not sound defensive when pushback is mild
- - Do not over-thank supportive comments
- - Do not dodge direct questions with vague language
- - Do not turn short replies into mini essays
- - Do not use the same stock response across a batch of comments
- - Do not escalate hostile or off-topic threads
- - Do not forget the original post context when it changes how the reply should land
-
- ## Validation Checklist
-
- - [ ] Reply matches the comment type
- - [ ] Reply uses the right tone mode
- - [ ] Reply length matches the depth of the comment
- - [ ] Reply accounts for the original post context when provided
- - [ ] Sentiment was handled appropriately
- - [ ] Tone sounds like Mick, not a generic customer-service voice
- - [ ] Final copy avoids emojis and em dashes
- - [ ] Reply is short enough for Facebook comments
- - [ ] Batch replies stay distinct when multiple comments are provided
- - [ ] Repeated questions can be collapsed into one pinned clarification when needed
- - [ ] Output is paste-ready without extra framing
-
- ## Help And Examples
-
- If the user is not sure how to use this skill, asks what it needs, or asks for examples:
-
- - Explain in plain language what this skill can do.
- - Tell the user the minimum input needed for a useful first pass.
- - Show the example prompts below.
- - Offer the fastest next prompt the user can send.
-
- Minimum useful input:
-
- - The comment or thread context plus the tone you want back.
-
- Example prompts:
-
- - `Use facebook-reply to answer this comment in a friendly but direct way.`
- - `Write three reply options to this Facebook comment, one warm, one short, and one more technical.`
- - `Show me how to use this skill when I have a full thread and need several replies.`
-
+ ---
+ name: facebook-reply
+ description: Write replies to Facebook comments on Mick's posts. Trigger on requests to reply to a Facebook comment, especially when the user starts with `fbr`.
+ version: 1.1.0
+ ---
+
+ # Facebook Reply
+
+ ## Intent
+
+ Write practical Facebook replies that match Mick's tone and respond to the actual comment context.
+
+ ## Do Not Use When
+
+ - Do not use for standalone Facebook posts, private messages, moderation decisions, or arguments that should be ignored.
+ - Required context is missing and cannot be reasonably inferred.
+ - A more specific skill in this repo is a better match.
+
+ Write replies to Facebook comments in Mick's direct, casual voice. Adjust the tone based on the comment type, sentiment, reply goal, and the context of the original post.
+
+ ## Trigger When
+
+ - The user starts with `fbr`
+ - The user asks for help replying to a Facebook comment
+ - The user provides a comment and wants a paste-ready response for Facebook
+ - The user wants replies for multiple comments on the same post
+ - The user wants one clarification reply for repeated questions in a thread
+
+ ## Workflow
+
+ 1. Treat everything after `fbr` as the reply context.
+ 2. If a post and comment are both provided, use both.
+ 3. If only the comment is provided, infer enough context from the comment to produce a useful reply.
+ 4. Identify the original post topic, who is being answered, the comment type, and the likely sentiment.
+ 5. Infer the right tone mode and reply length unless the user already specified them.
+ 6. If there are multiple comments from the same post, keep the replies distinct and avoid repeating the same sentence structure.
+ 7. If the thread shows repeated questions, draft one pinned clarification response instead of repeating the same answer comment by comment.
+ 8. Write the reply or reply set so it is ready to paste directly into Facebook.
+
+ ## Non-Negotiable Rules
+
+ - No emojis
+ - Avoid em dashes in final reply output
+ - Avoid quotation marks around words or phrases when they are not needed
+ - No hashtags in replies
+ - Never start with `Great question!` or `Thanks for asking!`
+ - First-person, casual, direct
+ - Most replies should be 2 to 4 sentences
+
+ ## Tone Modes
+
+ Infer one of these modes unless the user explicitly asks for a different one:
+
+ - `appreciative`: supportive comments, encouragement, or praise
+ - `clarifying`: direct questions, confusion, or requests for more detail
+ - `humorous`: light friendly comments where a short joke or playful line fits naturally
+ - `boundary setting`: criticism, bad-faith pokes, repeated off-topic comments, or replies that need a clean limit
+
+ ### Tone Guidance
+
+ - `appreciative`: warm but grounded. Thank them without overdoing it.
+ - `clarifying`: answer directly and keep the explanation tight.
+ - `humorous`: keep it dry or light. Do not get corny.
+ - `boundary setting`: stay calm, clear, and brief. Do not escalate.
+
+ ## Reply Length Modes
+
+ Infer one of these unless the user asks for a specific depth:
+
+ - `concise`: 1 to 2 short sentences
+ - `fuller`: 2 to 4 short sentences
+
+ Use `concise` for quick praise, obvious answers, or mild pushback. Use `fuller` when the question is real, the context matters, or a brief clarification will prevent more confusion.
+
+ ## Sentiment Tags
+
+ Classify the incoming comment before writing the reply:
+
+ - `supportive`
+ - `neutral`
+ - `skeptical`
+ - `critical`
+ - `off topic`
+ - `hostile`
+
+ Use the tag to decide how much warmth, detail, humor, or distance the reply should carry. Do not echo hostility back.
+
+ ## Tone By Comment Type
+
+ ### Positive Or Supportive
+
+ - Warm but grounded
+ - Acknowledge without gushing
+ - Add a concrete detail or next step when it fits
+
+ ### Question About A Project
+
+ - Answer the actual question directly
+ - Use specific details when known
+ - Offer more detail only if it feels natural
+
+ ### Skeptic Or Pushback
+
+ - Be confident without sounding defensive
+ - State the point clearly and move on
+ - Do not invite an argument
+
+ ### Criticism Or Hostile Pushback
+
+ - Answer only when a reply has value
+ - Keep the response short and controlled
+ - Set a boundary when the comment is clearly baiting or repetitive
+
+ ### Off Topic
+
+ - Redirect briefly when useful
+ - Skip extra context if the comment does not deserve it
+ - Do not let the thread drift away from the original post
+
+ ### General Or Ambiguous
+
+ - Match the comment's energy without mimicking it
+ - Keep the reply conversational and natural
+
+ ## Voice
+
+ Mick's replies should sound like a person, not a brand. Short sentences. Contractions. No motivational language. No canned gratitude phrases.
+
+ Avoid phrases such as:
+
+ - `journey`
+ - `transformative`
+ - `game-changer`
+ - `it means a lot`
+ - `so proud`
+ - `incredibly rewarding`
+ - `This taught me that`
+ - `The takeaway here is`
+ - `Great question!`
+ - `Thanks for asking!`
+ - `I appreciate your support!`
+
+ ## Context Memory Prompts
+
+ When context is available, account for:
+
+ - what the original Facebook post was about
+ - who is being answered if that matters to the tone
+ - whether the commenter is supportive, curious, skeptical, or trying to start a fight
+ - whether the same question has already been answered elsewhere in the thread
+
+ Use that context to make the reply feel like part of the real conversation instead of a generic standalone answer.
+
+ ## Difficult Comment Handling
+
+ For skepticism, criticism, and off-topic comments:
+
+ - answer the real point once if it is worth answering
+ - stay matter of fact instead of defensive
+ - do not overexplain to people who are clearly not asking in good faith
+ - if needed, end cleanly instead of leaving the door open for a circular argument
+
+ ## Batch Reply Mode
+
+ If the user provides multiple comments from the same post:
+
+ - write a separate reply for each comment
+ - vary the openings and sentence rhythm
+ - keep the replies aligned to the same original post context
+ - avoid reusing the same exact explanation unless the question is identical
+
+ ## Thread Summarization Mode
+
+ If several comments are repeating the same question or misunderstanding:
+
+ - detect the repeated theme
+ - draft one short pinned clarification reply
+ - make it broad enough to cover the repeated question without sounding canned
+
+ ## Output Rules
+
+ - Output only paste-ready reply text
+ - For a single comment, output only the reply
+ - For batch reply mode, separate each reply clearly with simple labels such as `Reply 1`, `Reply 2`, and `Pinned reply` when needed
+ - Do not add explanation around the replies unless the user explicitly asks for it
+
+ ## Examples
+
+ Supportive comment:
+
+ ```text
+ That leg press number is insane, great work!
+ ```
+
+ Good reply:
+
+ ```text
+ Appreciate it. Started at 465 three months ago so the improvement has been real. Still a lot of room to push.
+ ```
+
+ Question:
+
+ ```text
+ How are you powering the container home off grid?
+ ```
+
+ Good reply:
+
+ ```text
+ Solar with battery storage. Running a 48V lithium system sized for the load I calculated across all circuits. Still dialing in the generator backup for extended cloudy stretches.
+ ```
+
+ Pushback:
+
+ ```text
+ You're going to regret building that far out from civilization.
+ ```
+
+ Good reply:
+
+ ```text
+ Maybe. Property taxes are $400 a year and I own it outright. That trades off a lot.
+ ```
+
+ ## Output Format
+
+ Output only the reply text, ready to paste into Facebook. No labels, no explanation, and no extra commentary around the reply.
+
+ ## Constraints
+
+ - Do not sound defensive when pushback is mild
+ - Do not over-thank supportive comments
+ - Do not dodge direct questions with vague language
+ - Do not turn short replies into mini essays
+ - Do not use the same stock response across a batch of comments
+ - Do not escalate hostile or off-topic threads
+ - Do not forget the original post context when it changes how the reply should land
+
+ ## Validation Checklist
+
+ - [ ] Reply matches the comment type
+ - [ ] Reply uses the right tone mode
+ - [ ] Reply length matches the depth of the comment
+ - [ ] Reply accounts for the original post context when provided
+ - [ ] Sentiment was handled appropriately
+ - [ ] Tone sounds like Mick, not a generic customer-service voice
+ - [ ] Final copy avoids emojis and em dashes
+ - [ ] Reply is short enough for Facebook comments
+ - [ ] Batch replies stay distinct when multiple comments are provided
+ - [ ] Repeated questions can be collapsed into one pinned clarification when needed
+ - [ ] Output is paste-ready without extra framing
+
+ ## Help And Examples
+
+ If the user is not sure how to use this skill, asks what it needs, or asks for examples:
+
+ - Explain in plain language what this skill can do.
+ - Tell the user the minimum input needed for a useful first pass.
+ - Show the example prompts below.
+ - Offer the fastest next prompt the user can send.
+
+ Minimum useful input:
+
+ - The comment or thread context plus the tone you want back.
+
+ Example prompts:
+
+ - `Use facebook-reply to answer this comment in a friendly but direct way.`
+ - `Write three reply options to this Facebook comment, one warm, one short, and one more technical.`
+ - `Show me how to use this skill when I have a full thread and need several replies.`
+