Re-reading the same skills over and over again

What version of Codex CLI is running?

codex-cli 0.150.1

What subscription do you have?

Pro

Which model were you using?

GPT-5.6 Sol

What platform is your computer?

Linux

What terminal emulator and version are you using (if applicable)?

No response

Codex doctor report

What issue are you seeing?

This is a problem I've noticed for many months now. The agent will re-read the same skills almost every single turn. For example, if I have a frontend skill, and I ask the agent to make a change that affects the frontend in some way, it will read the frontend skill first.

Then, if I ask it to do some other change, it'll read the skill for that.

Then, if I ask it to do another frontend related change, it'll read the frontend skill again.

This is very wasteful for context usage. Is this a model issue, or a CLI bug?

What steps can reproduce the bug?

Create two different skills, ask the agent to do tasks related to the different areas the skill covers in its description, alternate between prompts affecting those areas and count the number of times each skill is read in a single session.

What is the expected behavior?

No response

Additional information

No response

Reproduce the reported behavior in a Linux Codex CLI session using two skills and alternating prompts, then count each skill's reads. Trace whether the repeated reads originate in Codex CLI or the model; done means the cause and expected behavior are established, with a regression check or clear explanation.