mcpField
A
autodev

autodev

github.com/boshu2Agent workflows45048 installsUpdated 1d ago
Repository

About autodev

autodev defines and validates a repository-local contract for autonomous development, including scope, validation, escalation, and stop conditions. It is for teams setting operating rules; separate drivers such as evolve run the development loop.

Key features

  1. 01Create and inspect development contracts
  2. 02Validate contract sections and rules
  3. 03Define mutable and immutable scope
  4. 04Set validation, escalation, and stop conditions
  5. 05Route work to evolve or rpi

Use cases

  1. 01Initialize an autonomous development policy
  2. 02Repair missing or invalid contract sections
  3. 03Check policy before running an autonomous loop

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/boshu2/agentops --skill autodev

After installing, type/autodevin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

skill.spec.json
references/
autodev-cli.feature
autodev.feature
scripts/
validate.sh

Details

From the project README

/autodev

/autodev manages the repo-local operational contract for autonomous development. It does not replace evolve or rpi.

Loop position

The config/intent layer the operating loop reads each tick — NOT a loop itself. /autodev defines and validates the contract declared in PROGRAM.md/AUTODEV.md — mutable scope, immutable scope, validation commands, escalation rules, stop conditions. The drivers (evolve and Factory, the daemon) consume that contract and run the loop; autodev does not run it. Loop discipline still applies under autonomy: no parallel wave without the wave-validity check; no slice closes without a passing test mapped to a Given/When/Then; capture goes through the promotion ratchet, not into a landfill.

  • PROGRAM.md or AUTODEV.md defines the contract: mutable scope, immutable scope, experiment unit, validation commands, decision policy, escalation rules, and stop conditions.
  • ao autodev creates, inspects, and validates that contract.
  • evolve runs the v2 autonomous improvement loop.
  • The operating loop runs one research -> plan -> implement -> validate lifecycle.

Routing

In Claude Code, /autodev hands work to /evolve or /rpi as skill invocations. Terminal wrappers remain for humans or non-skill runtimes; they are not the skill handoff default.

Use this split when the user asks whether the old evolve flow should become a new command or skill:

IntentAction
define or repair the repo-local autonomous policyuse /autodev and ao autodev
run the autonomous improvement loopinvoke /evolve
run one bounded lifecycleinvoke /rpi

PROGRAM.md takes precedence over AUTODEV.md. Treat AUTODEV.md as the compatibility alias.

Execution Steps

Step 1: Detect the contract
if [ -f PROGRAM.md ]; then
  PROGRAM_PATH=PROGRAM.md
elif [ -f AUTODEV.md ]; then
  PROGRAM_PATH=AUTODEV.md
else
  PROGRAM_PATH=
fi

If a contract exists, validate before using it:

ao autodev validate --json ${PROGRAM_PATH:+--file "$PROGRAM_PATH"}

If no contract exists and the user asked to initialize or define the loop, create one:

ao autodev init "<objective>"

Infer the objective from the user request when it is clear. Ask only when the objective cannot be discovered from repo context and inventing it would make the contract misleading.

Step 2: Repair or explain validation failures

When validation fails, inspect the missing fields and patch the program file if the user asked to create or fix the contract. Required sections:

  • Objective
  • Mutable Scope
  • Immutable Scope
  • Experiment Unit
  • Validation Commands
  • Decision Policy
  • Escalation Rules
  • Stop Conditions

Prefer narrow mutable scope and concrete validation commands. If the needed work crosses immutable scope, create or update a bead instead of silently widening the contract.

Step 3: Hand off to the loop

After ao autodev validate passes:

  • For one lifecycle, run the operating loop in-session; use NTM/Agent Mail or ao agent when the loop must run out of session.
  • For the repeated autonomous loop, invoke /evolve (skill-driven) or dispatch it via NTM for out-of-session runs.
  • If both PROGRAM.md and GOALS.md exist, GOALS.md is strategic fitness and PROGRAM.md is the operational execution layer.

Do not mark an autonomous cycle successful only because the main tests pass. The program validation bundle and stop conditions must also be satisfied.

Examples

User: turn this postmortem/analyze/plan/pre-mortem/implement/validate loop into
a v2 command.
Agent: Explain that `evolve` runs the loop, then create or validate
`PROGRAM.md` with `/autodev` so the loop has explicit scope and gates.
ao autodev init "Continuously improve AgentOps skills within explicit scope."
ao autodev validate
# then run the repeated loop: /evolve (skill-driven; /evolve CLI retired, ag-llni)

Troubleshooting

ProblemResponse
PROGRAM.md not foundRun ao autodev init "<objective>" when setup is requested.
validation reports missing sectionsPatch the missing required sections, then rerun ao autodev validate --json.
requested work is outside immutable scopeStop direct edits and create a bead or ask for an explicit contract change.
user asks "is this evolve?"Answer: autodev defines the loop contract; evolve runs the loop.

Reference Documents

Comments

Markdown supported