Autonomous

Skill mattjaikaran/meridian/skills/autonomous

Skill that gives AI coding agents persistent memory and quality enforcementFrom the repository description

Install

npx -y skills add mattjaikaran/meridian --skill autonomous

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

3 things to look at

skips confirmationTells the agent to proceed without asking first, 1 time: "Run discuss→plan→execute for each remaining phase without manual intervention".
1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
runs commandsInstructs the agent to run 4 commands, including `PYTHONPATH=$MERIDIAN_HOME uv run --project $MERIDIAN_HOME -- python -c "..."` and 3 more.

SKILL.md

1.5 KB, 389 tokens by cl100k_base, as published. Nobody here has run it

/meridian:autonomous — Hands-Free Execution

Run discuss→plan→execute for each remaining phase without manual intervention.

Arguments

--from <phase_id> — Start from this phase.
--to <phase_id> — Stop after this phase.
--only <phase_id> — Run a single phase autonomously.

Keywords

autonomous, auto, hands-free, unattended, run all, batch execute

Procedure

Plan the run — Determine which phases need processing.

PYTHONPATH=$MERIDIAN_HOME uv run --project $MERIDIAN_HOME -- python -c "
import json
from scripts.db import open_project
from scripts.autonomous import plan_autonomous_run
from scripts.state import get_status
with open_project('.') as conn:
    status = get_status(conn)
    ms = status.get('active_milestone')
    if not ms:
        print('ERROR: No active milestone')
    else:
        result = plan_autonomous_run(conn, ms['id'])
        print(json.dumps(result, default=str, indent=2))
"

For each phase, determine the next step and dispatch:
- discuss → Run /meridian:discuss --auto for the phase
- plan → Run /meridian:plan for the phase
- execute → Run /meridian:execute for the phase
- verify → Transition to reviewing
- complete → Skip (already done)
On failure, stop the loop and report which phase/step failed with error context. The state is resumable — running /meridian:autonomous again picks up where it left off.

Output

Show progress per phase: phase name, current step, result. On completion, show summary of phases processed.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.