Core Principle
We are not:
- pitching a tool
- asking for access
- convincing people
We are:
entering real workflows, adding value, and earning trust → then creating loops
Channel Strategy
1. GitHub Issues (Primary - 70%)
Why:
- highest intent
- real-time problems
- natural context
Goal:
turn issues into conversations → conversations into second runs
Step-by-step
-
Find active issues:
- bugs
- confusion
- dependency problems
- “why is this happening?” threads
-
First comment (No Pitch)
Structure:
- reference something specific
- suggest a plausible cause
- add value even if you disappear
Example:
"this might be related to X — looks like Y change didn’t propagate to Z module, seen similar issues when dependency graphs aren’t fully tracked"
- Wait for reply
If no reply:
If reply:
- engage naturally
- ask 1 clarifying question max
- Introduce Ryva (Soft)
"I’ve been experimenting with something that maps this kind of dependency automatically — can run it on this if useful"
- Show output
- clear
- structured
- specific to their issue
- Force Loop
Key line:
"this will likely change after your next merge — worth re-running"
OR
"there are still unresolved parts here, I’d check again after X"
2. Reddit (Secondary - 20%)
Why:
- good for conversations
- not good for compounding
Goal:
practice messaging + get initial users
Strategy
- only comment on high-signal posts
- do not pitch immediately
Structure:
"I’ve seen this happen a lot — how do you usually track what changed after merges?"
Then:
"I’ve been testing something that maps that automatically, happy to run it for you if useful"
Rule
Reddit is for:
- learning language
- testing hooks
Not for:
3. X / Twitter (Supporting - 10%)
Why:
- builds passive trust over time
Goal:
make people trust you before you DM them
What to post
- "ran this on a repo → here’s what it found"
- "most teams miss this after PR merges"
- "this is why standups exist (but shouldn’t)"
No fluff.
Just insights.
Trust Strategy (Critical)
We never:
- ask for repo access upfront
- push calls early
- overclaim ("this breaks prod")
We always:
- show value first
- use public repos when possible
- keep asks small
Loop Design (Most Important)
We are not optimizing for:
We are optimizing for:
second runs
Every interaction must end with:
- time trigger:
"check this again in 2 days"
OR
- event trigger:
"run this after your next PR"
Daily Execution
Daily targets
- GitHub Issues: 10–15 high-quality comments
- Reddit: 3–5 thoughtful comments
- X: 1 post (optional but ideal)
Tracking
For each interaction track:
- reply? (yes/no)
- used Ryva? (yes/no)
- came back? (yes/no)
What Success Looks Like
Not:
- viral growth
- tons of users
But:
1 user → 2nd run → 3rd run → expectation
That’s the seed of compounding.
What To Avoid
- jumping to calls too early
- pitching before context
- generic comments
- switching channels constantly
- building new features
North Star
"did a stranger come back and use this again without me pushing?"
If yes → double down
If no → fix loop
Final Rule
Do not chase:
- new ideas
- new features
- new channels
Until:
you get repeat usage from strangers
Everything else is distraction.