Notebook session ended
This session ended after eight hours.
Sign in again to continue from where you left off.
Notebook gate offline
The sign-in service is not responding right now.
Wait a minute and try again, or read what is already public from the hub.
Foundational walk through the eva project. Read top to bottom if you're new; no prior context assumed.
More about this project on the Eva hub.
Use when: scaffolding a new prompt under prompts/; auditing an unfamiliar prompt directory; wiring a guard or verifier into an existing prompt
Use when: authoring a new meta.yml; debugging eva doctor failures on required fields; preparing a prompt for skill export
Use when: debugging why a {{var}} did not substitute; deciding between --var, --var-file, and --case; writing a guard.sh that overlays variables
Use when: writing a pre-send validator that aborts bad inputs; writing a post-send checker that drives the verified counter; debugging why kick aborted before sending
Use when: computing prompt usage stats outside of eva show; debugging promotion gate counts; auditing performance medians for a prompt
Use when: promoting a prompt from draft to tested; promoting a prompt from tested to ready; diagnosing why eva promote refuses to bump status
Use when: adding a new case to eval.yml; choosing between assertions and a rubric judge; debugging a failing eva eval run
Use when: tuning meta.description so the right queries auto-load the skill; adding negative examples to eval.yml.triggering; diagnosing a should_match failure
Use when: diagnosing a doctor FAIL line; deciding whether a warning will block skill export; chasing an alias collision across prompts
Use when: regenerating the README catalog after adding a prompt; wiring a new shell verb into the kick-aliases.sh source; understanding which section of kick-aliases.sh is hand-edited
Use when: exporting a ready prompt as an Anthropic skill; debugging an export-skill blocked by doctor; understanding the SKILL.md frontmatter eva produces
Use when: fixing an 'id contains reserved skill token' failure; trimming meta.description to fit the 1024-char cap; deciding whether to split prompt.xml before export
Use when: diagnosing a missing-ruamel error from eva; running eva from a system Python instead of the project venv; moving the eva-hq checkout to a new machine
Use when: running the prompt catalog locally before publishing; adding a new template under site/prompts/; understanding which data source the catalog reads
Use when: writing the EVA tagline or homepage hero; explaining EVA to someone who builds with LLMs
Use when: qualifying inbound interest; deciding whether to feature EVA in a developer newsletter
Use when: writing the homepage hero copy; drafting a tweet or Show HN post for EVA
Use when: explaining why EVA does not have a launch campaign; deciding what counts as 'shipped' for EVA
Use when: responding to 'couldn't you do this in Obsidian?'; comparing EVA against Notion, GitHub gists, or a scratch directory
Use when: answering 'is there a paid tier?'; explaining EVA's sustainability story
Use when: framing why a single-user prompt catalog is a venture-shaped opportunity; answering 'why now' for personal AI tooling
Use when: sizing the addressable market for personal prompt tooling; answering 'how big can a single-user tool get?'
Use when: showing diligence-grade traction for a single-operator tool; answering 'what has actually been built?'
Use when: answering 'couldn't a notes app or GitHub do this?'; explaining what compounds as the catalog grows
Use when: explaining what the next round of capital actually buys; answering 'how do you go from a single user to a market?'