Feature flag cleanup planner

Skill sisodiabhumca/agent-skills/skills/feature-flag-cleanup-planner

Production-Ready Agent Skills : product analytics, growth experiments, CRM, research synthesis, postmortems, data contracts, SaaS spend, compliance, architecture maps, and LLM eval and many more.

Install

npx -y skills add sisodiabhumca/agent-skills --skill feature-flag-cleanup-planner

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

One thing to look at

0 stars0 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.

What its author says it does

Copied from the file, not written here

Vendor-neutral skill to prioritize feature flags for cleanup using simple heuristics and produce a deprecation plan.

SKILL.md

1.6 KB, as published. Nobody here has run it

When to invoke

You have a list of feature flags and want to reduce config debt safely.
You want to identify stale flags, flags permanently enabled, or flags with no recent evaluation.

Inputs needed

--flags: JSON file containing an array of flags with metadata.
Fields expected per flag (best-effort):
- key (string)
- created_at (ISO date)
- last_evaluated_at (ISO date, optional)
- owner (string, optional)
- default (on|off)
- environments: object like { "prod": {"state":"on|off", "percentage":0-100} }
- notes (string, optional)

Workflow

Parse and normalize the flag inventory.
Compute heuristics per flag:
- age in days
- days since last evaluation
- permanently on/off in prod
- risky rollout (prod percentage between 1 and 99)
Assign a recommended action:
- remove (stale + permanently on/off)
- migrate_to_config (long-lived flag acting like config)
- keep (recently evaluated or in active rollout)
Output a ranked plan with suggested steps and a checklist.

Output format

JSON:

summary
plan: list of {key, score, recommended_action, rationale, checklist} sorted by score descending

Guardrails

Never recommend removing flags in active rollout (1–99% in prod).
Prefer migrate_to_config if a flag is old and permanently on but likely used as a kill-switch.
Output must be explainable and deterministic.

Reference code

plan_cleanup.py