Viktor.comAn AI coworker that actually does the work
Creed
Your personal context file for every agent.
204 followers
Your personal context file for every agent.
204 followers
Personal context for your agents. Creed gives every connected AI a shared, up-to-date understanding of who you are, how you work, and what matters.
Creed
The 'goes stale the second I update it in one place' problem is the real test here: a shared context file only helps if every connected agent reads the current version, not a cached copy from last week. When I edit my Creed, does each agent pull it live at session start, or does it sync on a schedule and risk serving a stale snapshot mid-task? And is the source of truth hosted on your side, or a file I own locally that tools point at?
Connor, the thing I re-explain most isn't who I am — it's my domain. I do research-heavy work, and every new agent has to be re-taught my field's vocabulary, which sources I actually trust, and the shorthand I use, before it's useful. So the tool I'd most want plugged in first is my knowledge base (Obsidian / Zotero / NotebookLM), because that's where the real "how I think" lives, not my chat history. One design question: do you separate stable identity context (role, preferences) from fast-moving project context? The first I set once; the second changes weekly, and I'd worry a single flat file quietly blurs the two together.
This scratches a real itch. I've hand-copied the same claude md into every new agent and watched it go stale the second I changed one copy, so one portable file every tool reads is the obvious right shape. Making it something you own and edit rather than a black-box profile is the part I like most.
One thing I'd want before wiring it into everything is read scope. My coding agent, an email assistant, and some random marketing tool don't all need the same slice of "who I am." Wondering if creed lets me control which connected agent sees which part of the file, or whether every tool I connect gets the whole identity doc. The least-privilege version of this is a lot easier to trust than all or nothing
I've been solving this with a hand-maintained AGENTS.md per repo and the drift problem is real — it goes stale the moment my workflow changes. How does Creed handle updates: do I edit the file directly, or does it learn from usage over time? Also curious how you keep one shared context file from bloating once every connected agent starts contributing to it.
What I re-explain constantly isn't identity, it's rules, I keep a project guidelines file (folder structure, naming, "don't put business logic in components") for exactly this reason. But the failure mode I hit isn't staleness, it's silent drift: the agent has the file in context and still doesn't follow it three files deep into a task, and nothing flags that it happened. Does Creed have any way to verify an agent actually adhered to the context vs. just having it available? That gap matters more to me than sync timing.
Went through the site before commenting. The MCP pull is the right delivery choice, since a file fetched before the answer rather than pasted into a prompt is what makes this work across tools instead of inside one.
Andrew asked above whether the agent reads all of the file as it grows. I think the prior question is whether it ever shrinks. Both diffs in your own screenshots add lines and remove none, +18 and +9 with a zero on the other side, and that is the shape every memory product drifts into, because agents are good at noticing what to add and have no instinct for what stopped being true. Accept a year of those and it is a long preamble pulled before every answer, mostly stale, all of it paid for on every call.
So the question: does anything ever propose a deletion? A diff that says this line has not been true for four months, here is why, is the harder thing to build and the thing that keeps the file worth reading.
The other one is precedence. Your own example says ship to production Monday through Thursday. If my Creed says that and a repo's own context file says otherwise, I want to know which one the agent is meant to believe before I connect it to anything real.