Model behaviour issues on IntelliJ
Issues the maintainers labelled both platform:intellij and area:model.
140 issues · 10 open · 130 resolved (93%) · first seen Mar 5, 2026
Is this getting better or worse?
This class of problem is still growing. 112 new reports in the last 90 days vs 22 in the 90 before — +409%. The open backlog peaked at 10 in 2026-10 and sits at 10 today.
Derived by counting each matching issue from its creation date until it closed. First-party
computation over the anthropics/claude-code tracker — the trend is not published
anywhere upstream.
Which builds report this
43 of these reports name the Claude Code build they were running, spanning 38 releases. Heaviest reporters:
- Claude Code v2.1.295
- Claude Code v2.1.260
- Claude Code v2.1.259
- Claude Code v2.1.233
- Claude Code v2.1.208
- Claude Code v2.1.290
- Claude Code v2.1.286
- Claude Code v2.1.276
Report volume, not a defect count: an issue can mention a build without that build having caused the failure. Versions come from the version string in the bug-report template, matched against real published tags.
Do these get fixed?
Median time to close is 0 days across 130 closures. Of the 130 closures with a recorded reason, 4% were closed as completed and 125 as not-planned or duplicate. The most recent completed fix landed in 2026-07.
Matching issues 1–100
Ranked by community engagement (reactions weighted over comments).
- [Bug] Still reproducing on 2.1.186: assistant hallucinates fake user/system turns with stop_reason=null (re-report of #64791, wrongly auto-closed)
- [BUG] Claude doesn't respect hard rules to not NEVER use python for its script
- [BUG] the model produced a very long repetitive text output ("All files are read. Making all changes now.") in a loop without calling any tools, after completing a batch of parallel Read tool calls.
- [Bug] Security audit request flagged as inappropriate despite user ownership verification
- [Bug] Claude regressing on file change detection and autonomous file inspection
- [Bug] Claude Code generating responses to unrelated prompts
- [Bug] Model ignores user instructions and reverts to default behavior
- [BUG] claude-opus-4-7[1m] drifts from auto-loaded CLAUDE.md and memory notes across long sessions
- [Bug] Claude 4.7 ignores established project constraints and over-generates code/docs
- [Bug] Unresolved issues marked as pre-existing instead of addressing regressions
- [Bug] Claude ignores project context and memories, requires constant redirection
- [Bug] Fable 5 incorrectly flags prompts and tasks
- [Bug] Content Safeguard Incorrectly Flags Legitimate Debugging Activity
- [Bug] Unexpected model switching from Fable 5 to Opus during task execution
- [Bug] Incorrect content policy flag for non-security/biology topics
- [Bug] Instruction rejected without explanation when analyzing specification files
- [BUG] Claude Code falsely claims it can accept and read pasted screenshots in the terminal, even though terminals do not support image input.
- [Bug] Claude Code unable to perform basic drone control station operations
- [Bug] Claude Code consuming excessive tokens without producing code output
- [BUG] Claude Code wastes wall-clock time polling for background tasks that already completed
- [Bug] Unexpected language switch to Dutch in API responses
- ---
- [BUG] Claude cannot reliably output non-ASCII Unicode characters (e.g. U+2019 RIGHT SINGLE QUOTATION MARK) in file edits
- [Bug] ShopifyQL Query Knowledge Data Consistently Incorrect
- [Bug] Claude ignores explicit auto-invocation rules in context
- [Bug] Model falsely reports successful command execution when commands fail
- [Bug] Anthropic API Error: Safeguard False Positive on Legitimate Code Task
- [Bug] Claude Code not following explicit instructions and making incorrect assumptions
- [Bug] Model creates Jira tickets without explicit affirmative approval despite standing instructions
- [Bug] Agent performs unsolicited tasks without user confirmation, wasting tokens
- [Bug] Auto mode context handling misaligned with inquisitive queries
- [Bug] Unexpected automatic model switching to Opus
- [Bug] Excessive token usage when adding class fields - reads unrelated classes
- [BUG]
- [Bug] Anthropic API Error: Request Flagged for Security Policy Violation
- [Bug] Claude 3.5 Opus model performance degradation
- [Bug] Overly sensitive security classification blocking legitimate secret management operations
- [BUG] Auto-generated terminal tab title uses wrong language
- [Bug] Claude 3.5 Sonnet safeguard triggering unexpectedly, forcing fallback to Opus model
- [Feature Request] Add support for GreyHack game code generation with safety exceptions
- [Bug] Model Misattributes Slow Startup to Font Cache, Claims Deliberate Sabotage
- it allways damage code that was allready tested, no way to avoid damage
- [Bug] Content Policy Block: Missile.cpp Classification Triggering Harms Filter
- [Feature Request] UI testing agent should capture screenshot before attempting interactions
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author
- Withdrawn by the author