Capability · Feature flags
Every Feature Flag,
Finally on the Map
Flags ship fast and linger forever: forks in your code that no dashboard tracks once the launch is over. Code Swan records every flag your code reads, what each one switches, where it is read, and who it could affect, with evidence for every statement. Your flag provider knows which flags are on; now something knows what they do.
What the catalog knows about each flag
- What each flag actually switches: both sides, in plain terms
- Every place that reads it, across all your components and repositories
- Who a flag could affect, all the way to the APIs and events behind it
- Flags from your provider and homegrown flags alike, in one inventory
- Evidence for every statement, so nothing is taken on faith
- Honest scope: Code Swan reads your code, not your flag provider, so it never claims to know whether a flag is on
Why it matters
Flags are the fastest way to ship and the slowest thing to clean up. The cost is not the flag itself. It is every engineer who has to reconstruct what it does, one incident, review, and support ticket at a time.
Retire Flag Debt With Evidence
Every stale flag is a fork your team maintains, tests around, and trips over in incidents. Ask which flags look ready to remove and get candidates backed by evidence: where each one is read, what each side does, and what disappears when it goes. Cleanup stops being archaeology.
Explain Behaviour Differences in Minutes
"Why does this work for every customer except one?" is a flag question more often than anyone admits. Trace the flag from the symptom to the fork in the code, see both paths and who sits behind each, and stop reading source for an afternoon to answer a support ticket.
Review and Test the Path That Will Run
Flags are visible everywhere your team already looks: on the architecture, in design documents, and on every component they touch. Reviews and test plans cover both sides of the fork, not just the one the author remembered.
One Inventory Across Every Provider
Flags from your provider and your own homegrown helpers land in the same catalog. Filter by provider, lifetime, owner, or who a flag could affect, and share any view as a link. Nobody maintains the spreadsheet, because there is no spreadsheet.
Ask your AI assistant, get flag-aware answers
Flag intelligence is part of the same catalog your AI tools already query over MCP, so the questions your team actually has are now questions it can just ask:
- "What does this flag switch?", answered with both code paths and the evidence behind them
- "Which flags look ready to remove?", with cleanup candidates and where each one is still read
- "Why does this behave differently for one customer?", with the flags on the path and what each side does
- Reviews, test plans, and flow explanations take the flags on the path into account automatically
Frequently Asked Questions
Do you connect to our flag provider?
No. Code Swan reads how your code is written, so it never knows whether a flag is currently on, or for whom. Every flag answer says so and names what to check in your provider. What it gives you is the half of the picture your provider does not have: what each flag actually switches in the code.
Which flag systems are recognised?
Flags from your provider and flags read through your own homegrown helpers are both recorded, in the same inventory.
Where do flags show up?
Everywhere your team already looks: a dedicated Flags page you can filter and share, each flag’s own page, the components that read it, the architecture views, feature design documents, and your AI assistant.
How does this help us remove flags?
A flag is safe to remove when you know every place that reads it and what both sides do, which is exactly what the catalog records, with evidence. Ask your assistant which flags look ready to remove and you get candidates to verify against your provider, instead of a quarterly cleanup ticket nobody starts.
Does it work with the AI assistant?
Yes. Flag questions are answered like everything else in the catalog, and the assistant’s existing review, test planning, and flow mapping help takes flags into account, so a test plan covers both sides of a fork it would otherwise not know exists.
Know what every flag actually does
Connect your repositories and the next scan puts every feature flag on the map: what it switches, where it is read, and who it could affect.