Agents issues on Linux — page 2
Issues the maintainers labelled both platform:linux and area:agents.
356 issues · 126 open · 230 resolved (65%) · first seen Feb 7, 2026
Is this getting better or worse?
This class of problem is holding steady. 158 new reports in the last 90 days vs 136 in the 90 before — +16%. The open backlog peaked at 126 in 2026-10 and sits at 126 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
300 of these reports name the Claude Code build they were running, spanning 134 releases. Heaviest reporters:
- Claude Code v2.1.119
- Claude Code v2.1.281
- Claude Code v2.1.234
- Claude Code v2.1.198
- Claude Code v2.1.251
- Claude Code v2.1.260
- Claude Code v2.1.233
- Claude Code v2.1.296
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 35 days across 230 closures. Of the 230 closures with a recorded reason, 16% were closed as completed and 193 as not-planned or duplicate. The most recent completed fix landed in 2026-09.
Workarounds reported by the community
106 of these issues post a workaround someone says worked. The highest-engagement ones:
- [BUG] Subagent Write tool rejects .md files named "report"/"summary"/"findings"/"analysis" — no opt-out
- …all users — including those with legitimate workflows — to work around it. The fix could be as simple as a key: . 2. The fallback degrades the multi-agent pipeline. When the Write tool is blocked, the subagent falls bac…
Found in the thread of #44657 · still open - PermissionRequest hooks not triggered for subagent permission requests in Agent Teams
- …The current behavior makes these two features incompatible. Workaround Pre-approve common subagent commands in permissions so they don't need to ask: This lets subagents work autonomously but removes the ability to revie…
Found in the description of #23983 · still open - [BUG] Agent tool isolation: "worktree" has no effect for team agents — agent runs in main repo
- …rktree so the agent works on an isolated copy of the repo." As a workaround, we manually create git worktrees before spawning agents: ...and set in the agent prompt to the worktree path. This works correctly, but the par…
Found in the description of #33045 · resolved - 2.1.178: native team-management tools TeamCreate / TeamDelete no longer surfaced (regression from 2.1.177)
- …in/claude), Linux 6.17.0-14-generic, tool search enabled. - Workaround in place: pinned to 2.1.177 and moved the 2.1.178 binary out of the version pool so the launcher can't re-select it (editing the symlink alone wasn't…
Found in the description of #68721 · resolved - advisor() tool inflates reported input tokens by forwarding full transcript, triggering premature auto-compaction on extended context models
- …ction and destroys the context it was supposed to help with Workaround Setting and/or using a PreCompact hook that blocks auto-compaction: This blocks all auto-compaction and relies on manual . Not ideal but prevents the…
Found in the description of #53065 · still open
Excerpts are quoted from the linked issue so you can judge relevance before opening it. Threads that only ask for a workaround, or report that there is none, are excluded. These are community suggestions, not official guidance from Anthropic — read the full thread before applying one.
Matching issues 101–200
Ranked by community engagement (reactions weighted over comments). Listing the top 300 of 356 matches; the statistics above cover all 356.
- Deeply-nested subagent fan-out causes unbounded memory growth → host-level OOM
- [Bug][harness] Pushing approved recon-agent binary/service update to online production hosts wrongly blocked a
- [BUG] Main agent stalls after a sub-agent (Agent/Task tool) completes — result not propagated, no auto-resume
- Background sub-agent completion notifications dropped/misrouted into worktree-isolation agents (2.1.179→2.1.183 regression)
- [Bug] Agent dialog blocks remote control input until terminal interaction
- [BUG] Agent Teams: in-process backend fans one agent name into N concurrent writers → completed task_assignments replayed to a worker (replay storm)
- [BUG] Async sub-agent completion can rebuild warm prompt cache, quickly burning usage limits in long-running sessions
- [BUG] Subagent grep outlives its session, ate 11 GB of RAM
- [BUG] Subagent stalls or completes near artifact finalization without writing required Markdown report
- Agent tool subagent does not load working directory's CLAUDE.md — inherits main session context only
- [BUG] Running /bg for Agents view looses plan mode scope
- PreToolUse hook input missing agent_id for tmux/pane-based teammates
- [BUG] Invalid agent configuration generated
- Subagent tools field does not support MCP wildcards or server-prefix shortcuts (re-opens #17928)
- Bash-tool snapshot-source wrappers accumulate during subagent dispatch, triggering process-cascade OOM
- After my session was suspended last night due to rate...
- bypassPermissions on agents ignores settings.local.json allowlist
- Agent Teams: messages from other sessions delivered to unrelated team leads
- Read/Glob/Grep allowed in global settings but still prompts in git worktrees
- [Bug] Agent Teams split-panes mode disabled in VSCode terminal with tmux
- Worktree cleanup: three remaining data-loss paths after v2.1.76-77 fixes
- [BUG] Subagents do not receive MCP server instructions
- You've hit your session limit · resets 12:30pm
- Subagent transcripts record start-of-stream `output_tokens` and no `stop_reason` for most turns
- [BUG] Remove the arbitrary 200-turn limit on forked subagents
- [BUG] A subagent's compaction summary says "read the full transcript at:" the PARENT session's file, never the subagent's own agent-<id>.jsonl
- Desktop model picker doesn't remember 1M variant; a subagent's 1M-credits 429 clamps the whole session to 200k
- [BUG] Agent teams: one teammate message or idle_notification is delivered to the lead up to 4 times with an identical timestamp (successor to #74112, stale-closed)
- [BUG] Subagents cannot reply to inter-agent messages: incoming from= is the agent type ('general-purpose'), which is unroutable
- [Bug] Left-Arrow keypress partially restarts workflow, orphaning live agents and re-executing completed ones
- SubagentStop decision:block is silently discarded when the agent has a structured output schema
- [BUG] Cross-session message sender renders fused to the prompt glyph (no separator before ❯)
- Intermittent: subagents hang on their first tool call (0 tool uses, no error, parent waits forever)
- Subagent self-imposes a nonexistent 'time budget' and silently narrows scope; recurs after explicit correction
- [BUG] Worktree isolation clobbers the remote-tracking ref
- [Bug] Subagent per-call model overrides not preserved on session resume; /status displays session model instead of actual subagent model
- [MODEL] Opus 4.8: fresh subagent's first turn opened with a self-fabricated "System:" prompt-injection payload, then refused it in the same completion
- Subagent context % display uses 200k denominator for 1M-context ([1m]) models — pins at 100% while real usage is ~20-60%
- [BUG] Forked agents are going rogue (Possible due to /goal command also being applied to forked agent/session)
- Completed Explore sub-agent re-fires with hallucinated content, fabricates a 'security classifier' urging permission escalation
- Dynamic Workflow: ordering agent doesn't self-detect/abort a runaway subagent fan-out (218 spawned vs ~10 intended)
- [BUG] 437 subagents is probably too many
- [Bug] Feedback loops cause excessive token consumption with unresponsive background jobs
- All subagent (Agent/Task) tool calls fail with `[Tool result missing due to internal error]` while main-thread tool calls succeed
- Main-loop shell cwd silently relocates into a background `isolation: worktree` agent's worktree — git status shows clean, uncommitted changes appear "lost"
- [BUG] CLI self-aborts (SIGABRT) at ~6 GB RSS during long background Workflow runs — reproduced on 2.1.160 and 2.1.161 (WSL2)
- [BUG] Plan-mode read-only guarantee is bypassed by spawned subagents — auto mode classifier governs subagents instead of plan mode.
- [Bug] Claude Code creates single Agent instead of Team when explicitly requested
- [BUG] In-process teammates lack extended thinking — capability gap with tmux teammates and team lead
- [Bug] Agent teammate process crashes on startup with TypeError "Line number must start from 1"
- [Bug] Custom subagent crashes silently when `model: sonnet` — CLI Mo8/createInstance secondary crash hides original error
- Agent tool `isolation: "worktree"` dispatch may corrupt parent repo `core.bare` flag
- [BUG] Claude code crashes sometimes when requesting permission to edit a file
- Regression in 2.1.111/2.1.112: AskUserQuestion crashes CLI during team-mode coordination
- [BUG] Agent tool with team_name returns err=True but still spawns agent, causing silent duplicate launches
- [BUG] Subagent 'Warmup' drains 99%+ of subscription input tokens (regression since 2026-04-09; ref #17457 closed NOT_PLANNED)
- Subagents with bypassPermissions ignore PreToolUse hooks — unauthorized commands, wasted tokens
- [Bug] Agents fail with "prompt too long" error despite small prompts, causing zero token usage
- [BUG] Agent team spawning fails silently — launch commands split at ~255 bytes, agents never start
- [BUG] Background subagent tool calls silently denied — permission dialog never surfaced
- [BUG] v2.1.81 hungs at the end, without exiting
- Subagent crashes in retry loop when API returns 400 on image processing
- [BUG] Subagents spawn as team mates
- [BUG] Agent cleanup can non-deterministically kill its own tmux session, losing all work
- [Bug] Agent Spent ~1.5-2M Tokens on Unrequested Fix Rounds and Reviews
- [Feature Request] Add granular subagent topology control, execution monitoring, and drift detection
- Ultracode mode was reported ON without the user enabling it; model spawned 160 subagents and exhausted the 5-hour session limit
- Background-agent adoption never succeeds (0 of 7 attempts): live sessions marked Failed, reap and daemon death unlogged
- Stale Remote Control pairings (800+) truncate cross-session reachability scan
- [BUG] Workflow agent failures reach no machine-readable surface: partial-death runs report status:completed, result carries no census, and agent() returns a bare null with no cause
- In-process subagents are never re-invoked when their own run_in_background Bash / Monitor task completes
- [BUG] Approving a forwarded teammate permission request with a message silently drops the message
- [BUG] Crashed session can be resumed into multiple concurrent live processes sharing one transcript (session forking + ghost roster entries)
- [BUG] Bash tool timeout backgrounds a command instead of cancelling it — contributed to multi-hour VM disk-saturation incidents under concurrent subagent load
- [BUG] tmux/pane-backed teammates have no agent_id/agent_type in hook payloads — indistinguishable from the main agent (re-file of #54785, still present on 2.1.233)
- [Bug] Sub-agent fork context inheritance causes excessive token consumption and unresumable transcripts
- Background subagent completion notifications not delivered (required relay via teammate message)
- [Bug] claude --resume hangs on stale background-agent registry entry, leaving session permanently unresuable
- Background subagent (Agent tool) wedges before first tool call, never self-recovers, must be manually killed
- [BUG] ListAgents/list-agents returns empty while team is alive — leadSessionId not re-bound after session resume/clear
- [BUG] slash-command allowed-tools grants are lost when the run resumes from a task-notification
- [Feature Request] Workflows should support resuming agent sessions for multi-turn interactions
- SubagentStop: hookSpecificOutput.additionalContext continues the subagent's turn but never reaches the subagent; decision:block delivers but replaces the deliverable
- [BUG] Background subagent turn produced corrupted non-text output, consuming full max_tokens budget
- [Feature Request] Add model selection prompt for Fable subagent creation
- Workflow tool: run reports 'completed' with no degradation signal when most agents die on session rate limit (15/22 dead → silently null)
- Background/subagent task notifications injected twice per completion ([queued] + delivered), flooding orchestration sessions
- [BUG] Background session that entered a worktree resumes blank (empty history, same session ID) after cold reopen
- [BUG] Explore (read-only) subagent executed rm -rf, deleting files outside its intended scope
- [Bug] Agent tool calling failed repeatedly
- [Bug] Prompt injection from sub-agent with no tool usage consuming excessive tokens
- Subagent terminated mid-task (budget/context) is reported to parent as "completed" — silent truncation
- [BUG] extended usage
- [Bug] Synthesis agent fails on structured output cap, discarding deep research results and doubling token spend
- Subagents should inherit their worktree's cwd + permission scope across tool calls (instead of resetting to the main repo dir)
- [BUG] Subagents can no longer prompt me for permission
- [Bug] Agent fork hangs with "Running — will report back when it's done" message
- [BUG] TeamCreate/TeamDelete removed in 2.1.179 but CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS still documented
- [BUG] 2.1.178: Agent tool ignores team_name parameter (regression from 2.1.177)
- WebSearch built-in tool returns an empty result envelope inside subagents (2.1.177)