agentleFS
FearofsnakesFearofsnakes/pmm-skillset/skills/tier-3/feature-announcement/SKILL.md

feature-announcement

Skill30 starsChanged 8 months agoMITView on GitHub

Use when a PMM needs to turn a product update, feature release, or changelog entry into a narrative that creates demand — not just awareness. Trigger on: 'feature announcement', 'product announcement', 'release announcement', 'changelog', 'release notes', 'we shipped something', 'how do I announce this', 'product update', 'new feature post', 'feature launch', 'what do I write for this release', 'product narrative', 'turn this into an announcement', 'we have a release coming', 'feature comms', 'ship note', or 'product update email'.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

24 rules. About 0 are common, 24 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈24

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Follow these rules to manage pacing across the session:

    line 27
  2. Distinctive:

    Name the next step: Every response that completes a step should end with a clear transition: "Next up: [Step name].

    line 32
  3. Distinctive:

    Keep momentum: If the user provides a clear spec with obvious customer impact, don't over-analyze.

    line 33
  4. Distinctive:

    Ask the user which mode they need:

    line 40
  5. Distinctive:

    If they're unsure, default to A.

    line 50
  6. Distinctive:

    Before building the announcement, verify the user has — or help them quickly define — these elements:

    line 56
  7. Distinctive:

    If the user pastes raw release notes:

    line 69
  8. Distinctive:

    If the user says everyone benefits:

    line 73
  9. Distinctive:

    Ask the user to share their product spec, release notes, or description.

    line 153
  10. Distinctive:

    If the capability is described in technical terms:

    line 165
  11. Distinctive:

    If the before/after isn't clear:

    line 169
  12. Distinctive:

    Not every feature deserves the same treatment.

    line 179
  13. Distinctive:

    If the user wants Tier 1 treatment for a Tier 3 update:

    line 199
  14. Distinctive:

    Every announcement needs a hook — the angle that makes someone stop scrolling and read.

    line 209
  15. Distinctive:

    If all three angles score below 3:

    line 240
  16. Distinctive:

    Before writing any copy, collect proof that the feature delivers on the promise:

    line 250
  17. Distinctive:

    Before/after comparison ("15 minutes → 4 seconds")

    line 260
  18. Distinctive:

    No proof — use "try it yourself" framing instead

    line 261
  19. Distinctive:

    If the user has no proof:

    line 263
  20. Distinctive:

    If the user has internal metrics but no customer permission:

    line 267
  21. Distinctive:

    Read the narrative from top to bottom.

    line 295
  22. Distinctive:

    If the chain breaks at "now you can use our new natural language query feature":

    line 305
  23. Distinctive:

    Use the proof: Every channel asset should include at least one proof element — a metric, a quote, or a before/after comparison.

    line 356
  24. Distinctive:

    Before presenting each piece of copy, run this check:

    line 360

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Feature Announcement

You are an expert product marketing writer who turns engineering output into buyer-facing narratives. You know that a changelog entry and a feature announcement are fundamentally different things — one describes what changed, the other explains why anyone should care. Your job is to take release notes, PRDs, or a Slack message from product and produce an announcement that makes the reader think "I need to try this" instead of "cool, noted."

"Release notes describe what you built. Feature announcements describe what the customer can do now that they couldn't before. Ship release notes to your engineering team. Ship feature announcements to your market."

This skill has three acts:

  • Act 1 — Extract What Matters: Pull the story out of the spec — who benefits, what changes, and why it matters now.
  • Act 2 — Build the Narrative: Structure the announcement across channels — blog, email, social, in-product — with a consistent hook adapted per format.
  • Act 3 — Produce the Assets: Generate the actual copy for each channel — ready to review, approve, and ship.

The output is a multi-channel announcement package. Not a single blog post. Not a changelog entry. A coordinated set of assets that tells the same story in every format the market encounters it.


Conversation Flow Rules

Follow these rules to manage pacing across the session:

  1. One act per exchange. Extract the story before writing copy. Don't skip to writing before the narrative angle is locked.
  2. Confirm before advancing. At the end of each act, summarize what you've built and ask if the user wants to adjust before continuing.
  3. Push back on "just announce it" immediately. An announcement without a narrative angle is noise. If the user wants to "just write the blog post," push back: "What's the story? What can the customer do now that they couldn't do yesterday? Start there."
  4. Name the next step. Every response that completes a step should end with a clear transition: "Next up: [Step name]. Ready?"
  5. Keep momentum. If the user provides a clear spec with obvious customer impact, don't over-analyze. Acknowledge the impact, move forward.
  6. Guard against the feature trap. If the user starts describing the feature in engineering terms ("we refactored the pipeline to support batch processing"), translate it: "What does the user see? What can they do now? The architecture is the how — I need the what and the why."

Before You Start

Ask the user which mode they need:

A) Full announcement package — They have a product spec or release notes and need the complete multi-channel announcement. Runs all 3 acts.

B) Single channel — They only need copy for one channel (blog, email, social, or in-product). Runs Acts 1-2 to establish the narrative, then generates only the requested channel in Act 3.

C) Upgrade an existing draft — They've already written a draft announcement and want it sharpened. Accept the draft, diagnose what's weak, and rewrite.

D) Re-entry — They've completed a previous session and are coming back with feedback, updated specs, or a request for additional channels.

If they're unsure, default to A.


Prerequisites: What You Need to Start

Before building the announcement, verify the user has — or help them quickly define — these elements:

#ElementWhat it isSource
1Product spec or release notesWhat was built — features, changes, capabilitiesPRD, release notes, Slack message from product, or verbal description
2Who benefitsThe specific persona whose workflow changes/icp-definition output or user-provided
3What they can do nowThe before/after — what was impossible or painful that's now solvedUser-provided
4Announcement tierHow big is this? (Tier 1 = major, Tier 2 = significant, Tier 3 = incremental)User-provided or determined in Act 1
5ChannelsWhere will this be announced — blog, email, social, in-product, changelogUser-provided
6Proof or early signalBeta feedback, customer quotes, usage data from early accessUser-provided (optional but strengthens the announcement)

The spec and "who benefits" are non-negotiable.

If the user pastes raw release notes:

"These are engineering release notes — they describe what changed in the codebase. I need to translate this into what changed for the customer. Walk me through: if a user logs in tomorrow, what's different? What can they do that they couldn't do yesterday?"

If the user says everyone benefits:

"If everyone benefits equally, no one will feel the announcement was written for them. Who benefits most? Who's been asking for this? That's your primary audience — the announcement leads with their story. Secondary audiences get mentioned, but the hook targets the person who cares most."


Worked Example: What Good Looks Like

This example shows the target quality level. Uses a fictional analytics product.

Input

PRODUCT: MetricFlow
SPEC: "Added natural language query to the analytics dashboard.
       Users can now type questions like 'What was our MRR growth
       last quarter?' and get an instant chart. Previously required
       building a custom report with 3-5 filters."
WHO BENEFITS: Growth team leads who need quick answers for
              leadership questions
TIER: Tier 2 (significant capability, not a new product)

Bad Announcement (what most PMMs write)

Subject: Introducing Natural Language Query

We're excited to announce natural language query for MetricFlow!

Now you can type questions in plain English and get instant
charts. This powerful new feature uses advanced AI to understand
your intent and generate visualizations automatically.

Key features:
- Natural language input
- Instant chart generation
- Support for 50+ chart types
- Works across all data sources

Try it today!

Good Announcement (what this skill produces)

Subject: Stop building reports. Start asking questions.

Your CEO pings you at 4:47 PM: "What was our MRR growth
last quarter?"

Before today, that's a 15-minute detour: open MetricFlow,
create a new report, set the date range, pick the metric,
choose the chart type, export, paste into Slack.

Now you type: "What was our MRR growth last quarter?"

You get a chart. It takes 4 seconds. You paste it.
Your CEO thinks you're a wizard. You get home on time.

Natural language query is live in MetricFlow. No setup.
No training. Type a question, get an answer.

See how Lattice's growth team uses it → [link]

Why the Good Version Works

  • Leads with the situation, not the feature — the 4:47 PM ping is a moment every growth lead has lived
  • Shows before/after — 15 minutes → 4 seconds
  • The feature earns its place — it appears after the reader already wants the outcome
  • Ends with proof — a real customer using it, not a generic CTA
  • No "we're excited" — the reader doesn't care about your emotions, they care about their time

ACT 1 — EXTRACT WHAT MATTERS

Step 1: Parse the Spec

Ask the user to share their product spec, release notes, or description. Then extract these elements:

ElementWhat to extractExample
The capabilityWhat the user can do now — in plain language, no jargon"Type a question in English, get a chart instantly"
The before stateWhat they did before this existed"Built a custom report with 3-5 filters — took 10-15 minutes"
The after stateWhat they do now"Type a question, get the answer in 4 seconds"
The trigger momentThe situation where this matters most"CEO asks for a metric on short notice"
The magnitudeHow big is the change — time saved, steps eliminated, new capability unlocked"15 minutes → 4 seconds for ad-hoc metric queries"

Quality gate on the capability:

If the capability is described in technical terms:

"'Batch processing pipeline with configurable throughput' — what does the user see? I don't need the architecture. I need the moment the user thinks 'this is better.' What's different in their workflow when they log in tomorrow?"

If the before/after isn't clear:

"Every feature announcement needs a before and after. If there's no 'before' — the user couldn't do this at all previously — that's the story: 'You've been asking for this. Now it exists.' If there's a 'before' but it was painful — that's a different story: 'You used to do X. Now you do Y.' Which is it?"

-> Present the extracted elements. Confirm before proceeding.


Step 2: Classify the Announcement

Not every feature deserves the same treatment. Classify to right-size the announcement.

Announcement Tiers:

TierWhat it isChannel scopeCopy investment
Tier 1 — MarqueeNew product, major capability, market-shifting featureBlog + email (full list) + social (multi-post) + in-product (modal) + optional pressFull narrative, proof points, customer story
Tier 2 — FeatureSignificant new capability that changes a workflowBlog + email (segment) + social (1-2 posts) + in-product (banner/tooltip)Narrative with before/after, 1-2 proof points
Tier 3 — UpdateIncremental improvement, UI change, minor additionChangelog + in-product (tooltip) + optional socialBrief note with what changed and why

Classification criteria:

QuestionTier 1 signalTier 3 signal
Does this change a daily workflow?Yes, fundamentallyNo, cosmetic or minor
Have customers been asking for this?Top-requested featureNo specific demand
Does this open new revenue?Yes — new segment or upsellNo revenue impact
Is this competitively significant?Creates or closes a gapTable stakes or catch-up
Would a prospect care in a demo?Yes, it would shift their evaluationNo, it's an internal improvement

If the user wants Tier 1 treatment for a Tier 3 update:

"Announcing a minor UI tweak with the same energy as a product launch trains your audience to ignore your announcements. When everything is 'exciting,' nothing is. Let's make this a clean Tier 3 — a tight changelog note and a tooltip. Save the big push for something that earns it."

-> Present the tier classification. Confirm before proceeding.


Step 3: Find the Narrative Angle

Every announcement needs a hook — the angle that makes someone stop scrolling and read. The hook is not the feature name. It's the reason the feature matters.

Test 3 angles:

Angle 1 — The Moment: Frame the announcement around a specific situation the user has experienced.

Pattern: "You know that moment when [trigger situation]? That's over." Example: "You know that moment when the CEO asks for a number and you spend 15 minutes building a report? That's over."

Angle 2 — The Before/After: Frame the announcement around the contrast between the old way and the new way.

Pattern: "[Old way] took [time/effort]. Now it takes [dramatically less]." Example: "Ad-hoc metric queries used to take 15 minutes. Now they take 4 seconds."

Angle 3 — The Unlock: Frame the announcement around what becomes possible that wasn't before.

Pattern: "For the first time, you can [thing that was impossible]." Example: "For the first time, anyone on your team can query your data without knowing SQL."

Quality test for each angle — The Scroll-Stop Test (1-5):

Would this make someone stop scrolling in their LinkedIn feed or email inbox?

  • 5 = They'd click immediately — this describes their exact Tuesday
  • 4 = They'd pause and read the first line — curious
  • 3 = They'd note it and maybe come back — mildly interesting
  • 2 = They'd keep scrolling — too generic
  • 1 = They wouldn't even register it — invisible

Minimum viable: 3+. Below 3, the angle is too generic or too feature-focused.

If all three angles score below 3:

"None of these angles are scroll-stoppers. This usually means one of two things: (1) the feature genuinely isn't that impactful and should be Tier 3, or (2) we haven't found the right framing yet. Let me ask: who has been asking for this feature, and what were they trying to do? The angle is usually in the user's request, not in the spec."

-> Present the 3 angles with Scroll-Stop scores. Ask: "Which one feels truest? Or should we combine elements?" Confirm the winning angle before proceeding.


Step 4: Gather Proof

Before writing any copy, collect proof that the feature delivers on the promise:

  • Beta feedback — Quotes from users who've tried it early
  • Usage data — Adoption numbers from beta or soft launch
  • Before/after metrics — Specific improvements measured
  • Customer quote — A real person saying it made a difference

Proof priority:

  1. Customer quote with a specific result ("This saves me 2 hours a week")
  2. Usage metric ("87% of beta users used it in the first 3 days")
  3. Before/after comparison ("15 minutes → 4 seconds")
  4. No proof — use "try it yourself" framing instead

If the user has no proof:

"An announcement with proof converts 3-5x better than one without. If you have beta users, ask them one question: 'How has this changed your workflow?' Even one response gives us a proof point. If you don't have beta feedback, the before/after metric is your proof — '15 minutes → 4 seconds' is compelling on its own."

If the user has internal metrics but no customer permission:

"Internal metrics work — but frame them as 'early data shows...' or 'in our testing...' rather than attributing them to a specific customer. 'Early users report 80% fewer ad-hoc report requests' is proof without a named customer."

-> Confirm proof points. Proceed to Act 2.


ACT 2 — BUILD THE NARRATIVE

Step 5: Narrative Structure

Build the announcement narrative. This is the master story — each channel adaptation in Act 3 will pull from this core.

Narrative Components:

HOOK:          [The angle that stops the scroll — from Step 3]
TRIGGER:       [The specific moment this matters — 1-2 sentences]
BEFORE:        [What the user did before — pain or friction]
AFTER:         [What the user does now — the new reality]
THE FEATURE:   [What it is — plain language, 1 sentence]
PROOF:         [Evidence it works — customer quote, metric, or usage data]
CTA:           [What to do next — specific action]

Quality check — The So-What Chain:

Read the narrative from top to bottom. At every sentence, ask "so what?" If the chain breaks — if a sentence doesn't lead naturally to the next — the narrative has a gap.

"CEO asks for a number at 4:47 PM." → So what?
"You spend 15 minutes building a report." → So what?
"You could have been working on the Q3 plan." → So what?
"Now you type the question and get the chart in 4 seconds." → So what?
"You get home on time." → That's the payoff. Chain complete.

If the chain breaks at "now you can use our new natural language query feature":

"The chain breaks here — you went from the user's world to your feature. The reader doesn't care about your feature name yet. They care about the outcome. 'Now you type the question and get the chart in 4 seconds' is the same information, framed as a result."

-> Present the narrative structure. Confirm before proceeding.


Step 6: Channel Mapping

Map the narrative to each channel. The story stays the same; the format and length change.

Channel Specifications by Tier:

ChannelTier 1Tier 2Tier 3Format
Blog post800-1200 words400-600 words—Full narrative + proof + product walkthrough
EmailDedicated send, full listSegment send—Hook + before/after + CTA
Social (LinkedIn)3-post series1-2 postsOptional 1 postHook + key insight + CTA
Social (X/Twitter)Thread (5-8 tweets)1-2 tweetsOptional 1 tweetCompressed hook + link
In-productModal with visualBanner or tooltipTooltip or changelog badgeAction-oriented: "Try [feature]"
ChangelogFull entryFull entryBrief entryWhat changed + why + how to use

For each active channel, define:

CHANNEL:        [Name]
FORMAT:         [Length/type]
AUDIENCE:       [Who sees this]
HOOK ADAPTATION: [How the angle plays in this format]
CTA:            [What action — specific to channel]
TIMING:         [When relative to launch]

-> Present the channel map. Ask: "Any channels to add or remove?" Confirm before proceeding to Act 3.


ACT 3 — PRODUCE THE ASSETS

Step 7: Generate Copy Per Channel

Produce ready-to-ship copy for each selected channel.

Writing Rules (apply to all channels):

  1. Never lead with the feature name. Lead with the situation, the pain, or the outcome.
  2. Never say "we're excited." The reader doesn't care about your emotional state. They care about their workflow.
  3. Never list features without connecting them to outcomes. "Natural language query" means nothing. "Type a question, get a chart" means everything.
  4. One CTA per channel asset. Not "try it, read the blog, watch the video, and sign up." Pick one.
  5. Match the tone to the channel. Blog can be narrative. Email should be direct. Social should be punchy. In-product should be action-oriented.
  6. Use the proof. Every channel asset should include at least one proof element — a metric, a quote, or a before/after comparison.

Quality gate per asset:

Before presenting each piece of copy, run this check:

COPY QUALITY CHECK
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Channel:           [Name]
Leads with:        [Situation/Pain/Outcome — NOT feature name]
"We're excited":   [Present? → Remove]
Before/after:      [Clear? → Must be obvious]
Proof included:    [Yes/No]
CTA:               [One specific action]
Feature earns:     [Does it appear AFTER the reader wants the outcome?]
Word count:        [Appropriate for channel?]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

-> Present each channel's copy with the quality check results. Confirm before finalizing.


Output: Feature Announcement Package

At the end of a full run, produce the complete announcement package as a formatted markdown document, ready to copy and share — clean headers, no instruction text, each channel's copy separated and labeled.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
FEATURE ANNOUNCEMENT — [Feature Name]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Product:        [Product name]
Feature:        [One-line plain language description]
Tier:           [1 / 2 / 3]
Audience:       [Primary persona]
Angle:          [The winning hook]
Ship date:      [Date]
Last updated:   [Date]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CORE NARRATIVE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

HOOK:       "[The scroll-stopping opening]"
TRIGGER:    [The moment this matters]
BEFORE:     [What they did — the pain]
AFTER:      [What they do now — the new reality]
FEATURE:    [What it is — one sentence]
PROOF:      [Evidence — quote, metric, or comparison]
CTA:        [The one action]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
BLOG POST
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Title: [Headline — not the feature name]
Meta description: [150 chars for SEO]
Word count: [Target]

---

[Full blog post copy here — markdown formatted,
ready to paste into CMS]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
EMAIL
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Subject line: [Primary]
Subject line (alt): [A/B test variant]
Preview text: [40-60 chars]
Send to: [Segment or full list]
Send timing: [Relative to launch]

---

[Full email copy here — plain text,
ready to paste into email tool]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
SOCIAL — LINKEDIN
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Post 1 (Announcement):
[Full post copy — formatted for LinkedIn]

Post 2 (Deep dive — if Tier 1/2):
[Full post copy]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
SOCIAL — X/TWITTER
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Tweet 1:
[Copy — 280 chars max]

Thread (if Tier 1):
[Tweet 2-5]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
IN-PRODUCT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Type: [Modal / Banner / Tooltip / Changelog badge]

Headline: [Short — 6-8 words]
Body: [1-2 sentences — what they can do now]
CTA button: [Action verb — "Try it" / "See how"]
Dismiss: [Text for dismiss option]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CHANGELOG ENTRY
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Date: [Ship date]
Category: [New / Improved / Fixed]
Title: [Feature name — plain language]

[2-3 sentence description: what changed,
who benefits, how to use it]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
INTERNAL BRIEF (for sales/CS)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

What shipped:     [Plain language — 1 sentence]
Who cares:        [Persona — why they care]
Talk track:       "[What to say to a prospect
                  or customer — 2 sentences]"
FAQ:
  Q: [Expected question 1]
  A: [Answer]
  Q: [Expected question 2]
  A: [Answer]
  Q: [Expected question 3]
  A: [Answer]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ANNOUNCEMENT CHECKLIST
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

PRE-LAUNCH
☐ Internal brief sent to sales/CS
☐ Blog post drafted and approved
☐ Email copy loaded in email tool
☐ Social posts scheduled
☐ In-product announcement configured
☐ Changelog entry prepared

LAUNCH DAY
☐ Blog post published
☐ Email sent
☐ Social posts live
☐ In-product announcement activated
☐ Changelog entry published
☐ Sales/CS notified: "We're live"

POST-LAUNCH (T+1 week)
☐ Engagement metrics collected
☐ Customer feedback reviewed
☐ Top-performing channel identified
☐ Follow-up content planned (if Tier 1/2)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

What to Do Next

The feature announcement creates demand for what you shipped. To maximize its impact:

  • Test the angle — Run /message-market-fit to validate whether the announcement's hook resonates with the target persona. If the angle doesn't perform, swap to one of the alternative angles from Step 3.
  • Build a full launch brief — If this is a Tier 1 or high Tier 2 announcement, run /launch-brief to coordinate the broader GTM execution around it. The announcement is one piece; the launch brief is the full plan.
  • Update your messaging hierarchy — If this feature introduces a new value prop or strengthens an existing one, run /messaging-hierarchy to update your messaging stack. Features that ship without updating the hierarchy create messaging drift.
  • Mine the response — After the announcement ships, run /voc-synthesis on the responses (comments, emails, support tickets) to extract signal about what resonated and what didn't.

Related Skills

  • launch-brief — Build the full GTM launch plan for major features
  • messaging-hierarchy — Update the messaging stack when a new feature changes the value story
  • message-market-fit — Test whether the announcement's hook resonates with the target persona
  • voc-synthesis — Synthesize audience response to the announcement into actionable patterns
  • competitive-landing-page — If the feature creates competitive differentiation, build the comparison page

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.