e2e-session-test
Skill151 starsChanged 3 months agoMITView on GitHub
Use when smoke-testing cctop end to end — verifying a real coding-agent session is tracked and shows in the panel. Trigger on "smoke test sessions", "verify session tracking", "test the hook pipeline", "check cctop picks up a claude/codex/opencode/pi session", or "run the e2e test", even without the word "skill". Creates a real CLI agent session in a terminal, confirms cctop tracks it, opens the panel programmatically, and captures screenshot proof.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- cctop-restartSkill
- releaseSkill
- swiftui-debuggingSkill
- video-assetsSkill
- video-storyboardSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
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
- Distinctive: line 14
Verify a real session flows all the way through:
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.
Files
cctop end-to-end session test
Verify a real session flows all the way through:
agent in a terminal → cctop hooks → ~/.cctop/sessions/<pid>.json → panel cardA good smoke test after touching hooks, the session schema, or the panel.
Capabilities, not tools
Harness-agnostic — it needs three generic capabilities. Use your environment's equivalent; don't assume a specific tool:
- run shell commands
- capture a screenshot of the screen (or fall back to the
screencaptureCLI) - view an image you captured
Create the session with a CLI agent in a terminal
Drive a CLI agent (claude / codex / opencode / pi) in a terminal window —
it's a real session and one shell command starts it. Don't try to puppet a
desktop agent's GUI: if it's the app hosting you, screen capture hides its window
and you can't confirm input landed (see references/troubleshooting.md). Requires a
cctop build with the cctop:// URL scheme (ships with this skill).
Steps (run from the repo root)
# 1. Baseline — note the PIDs
.claude/skills/e2e-session-test/scripts/cctop-sessions.sh
# 2. Create a session — opens a terminal running the agent, waits for it to register
.claude/skills/e2e-session-test/scripts/launch-cli-session.sh claude ~/projects/cctop
# 3. Verify — a new PID entry with the right source (cc/codex/opencode/pi) appears
.claude/skills/e2e-session-test/scripts/cctop-sessions.sh
# 4. Summon the panel
open cctop://toggleThen capture a screenshot and confirm the new session's card is in the panel
(right status, source, project). Run open cctop://toggle again to close it — a
clean closed/open pair if you want before/after proof. For a shareable record,
write a small HTML log embedding the screenshots.
Two details the launcher handles, because they otherwise make this fail silently: a
login shell (so the agent's PATH, e.g. ~/.local/bin, resolves) and a
trusted directory (so the agent doesn't stall at a folder-trust prompt).
open cctop://toggle is used instead of a menu-bar click or hotkey because those
need extra OS permissions under automation.
PASS
- New tracked session within ~5s, correct
source. open cctop://toggleopens and closes the panel.- A screenshot shows the card.
On failure, grab the relevant ~/.cctop/logs/<session>.log entry and see
references/troubleshooting.md.
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
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.

