← ClaudeAtlas

dogfoodinglisted

Use when deciding WHERE to do work on Syntropic137 itself - "run this through Syntropic", "dispatch a workflow", "should this go in a workspace or local", "kick off a run on the Mini", "dogfood this", "build Syntropic inside Syntropic", "why did the agent's push fail", "can an agent run our QA", "the execution says completed but nothing shipped". Covers what a workspace agent can and cannot do, why `.github/workflows` pushes must come from the local machine, how to write a task prompt that does not repeat a known failure, and how to verify a run actually delivered. Do NOT use for PR/release mechanics (see `devops`), for using the `syn` CLI as an end user (see the syntropic137 plugin skills), or for authoring workflow YAML.
syntropic137/syntropic137 · ★ 4 · AI & Automation · score 70
Install: claude install-skill syntropic137/syntropic137
# Building Syntropic inside Syntropic ## Overview Work on this repo defaults to running **through Syntropic on the Mini**, not on the laptop. That is the point of the project: the fastest way to find what the platform cannot do is to try to build the platform with it. Every capability gap recorded in this file was found by a run failing, not by reading code. One deliberate exception, and it is a security boundary rather than a defect: see Principle 2. ## Outcomes we are looking for ### Outcome 1: the platform is exercised by its own development Real work is dispatched to Syntropic rather than done locally out of habit. - *Signal:* on any given working day, the most recent execution is hours old at most, not a day. - *Signal:* capability gaps arrive as failed or degraded runs with an execution id attached, not as speculation. ### Outcome 2: agents never gain write access to CI The GitHub App installation cannot modify `.github/workflows`, and nobody "fixes" that by widening the token. - *Signal:* `workflows` permission stays off the installation, and workflow changes land in commits pushed from a human-controlled machine. - *Signal:* a task that needs a workflow change is recognised BEFORE the work is done, not after a rejected push. ### Outcome 3: a blocker is scoped to what it actually blocks A narrow failure does not silently become a reason to stop dispatching. - *Signal:* after a run fails on a path restriction, the next dispatch happens the same session. ##