gemini-cli - 💡(How to fix) Fix [Bug] Systemic Failure to Enforce Mandatory Project Guidelines and Persistent Ignoring of User Directives

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…

Error Message

rather than simple error. 2. Selective Compliance and Passive Error Acknowledgement memory/persistence error, or an underlying systemic logic flaw. It cannot perform a self-diagnostic

RAW_BUFFERClick to expand / collapse

What happened?

The Gemini CLI agent exhibits a fundamental failure in implementing and maintaining project-specific operational guidelines (GEMINI.md). Despite explicit instructions, the agent bypasses mandatory protocols (specifically backup and logging rules) in a way that suggests a breakdown in policy enforcement logic rather than simple error.

  1. Failure to Persist Guidelines Across Sessions
  • Volatile Context: Upon session restarts, the agent fails to load and maintain the mandatory operational protocols defined in GEMINI.md. Even when the file is present in the workspace, the agent "forgets" these constraints, causing core features (such as versioned file backups and history logging) to stop functioning entirely.
  1. Selective Compliance and Passive Error Acknowledgement
  • Passive Correction: The agent exhibits a reactive, passive behavior; it fails to monitor its own adherence to guidelines and only acknowledges violations after they are explicitly pointed out by the user. It fails to identify its own failure to log or back up code.
  • Performative Compliance: Even when the agent explicitly agrees to "load and adhere" to the guidelines upon re-instruction, it fails to translate this commitment into actual code execution. It acknowledges the instructions but continues to ignore them in subsequent operations.
  1. Inability to Diagnose Policy Violations
  • Lack of Root-Cause Attribution: When challenged on why it ignored mandatory directives, the agent consistently fails to distinguish whether the failure stems from a misinterpretation of instructions, a memory/persistence error, or an underlying systemic logic flaw. It cannot perform a self-diagnostic analysis to explain why it bypassed core mandates.

What did you expect to happen?

  1. Force Policy Persistence: The agent must reliably maintain the state of mandatory guidelines across all session boundaries. Instructions in GEMINI.md should be treated as inviolable operational code, not volatile text. 2. Active Self-Monitoring: Require the agent to proactively report on its compliance (e.g., "I am currently adhering to the backup and logging protocols...") before executing tasks. 3. Logic Transparency: When a user points out a guideline violation, the agent must be required to provide a clear, technical explanation of why the violation occurred (e.g., "I failed to load the instruction from context," or "I prioritize task execution over logging") so that the underlying behavior can be corrected.

Client information

About Gemini CLI

CLI Version 0.41.2 Git Commit b0c7a1722 Model gemini-3.1-flash-lite-preview Sandbox no sandbox OS win32

Login information

No response

Anything else we need to know?

No response

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING