Four New Unnamed Feature Flags [In Development]

Four new feature flag constants have been added without any observable user-facing behavior yet:

tengu_amber_astrolabe (also: CLAUDE_CODE_AMBER_ASTROLABE env var)
tengu_bison_cairn (also: CLAUDE_CODE_BISON_CAIRN env var)
tengu_larch_cistern (also: CLAUDE_CODE_LARCH_CISTERN env var)
tengu_alder_wicket (also: CLAUDE_CODE_ALDER_WICKET env var)

These are infrastructure additions that will be wired to behavior in a future release.

Details

tengu_amber_astrolabe / CLAUDE_CODE_AMBER_ASTROLABE: accessor ekc() is checked in Fa_() alongside the existing max/autonomous mode gate (VMe(e)); when true, the autonomy system prompt section is injected even outside max mode, instructing Claude to proceed with reversible actions without asking "Want me to…?" permission questions and to stop only for destructive actions or genuine scope changes
tengu_bison_cairn / CLAUDE_CODE_BISON_CAIRN: accessor tkc() gates the "delivering_work_max" system prompt slot (el_); when true, a # Delivering work instruction block is added telling Claude to: act on what is in the request rather than speculation; deliver complete work under stated assumptions; explicitly report blocked sub-tasks; carry mid-task questions to the end of a turn that also delivers progress (rather than blocking); and treat user reaffirmation as a final decision to proceed
tengu_larch_cistern / CLAUDE_CODE_LARCH_CISTERN: accessor rkc() gates the "overcorrection" system prompt slot (tl_); when true, a # Corrections instruction block is added telling Claude to: only correct errors that would change the user's code, conclusions, or decisions; state corrections plainly and continue without apology, preamble, or rumination; not treat a follow-up question as a signal of prior error
tengu_alder_wicket / CLAUDE_CODE_ALDER_WICKET: accessor nkc() gates the "scope_fidelity" system prompt slot (rl_); when true, a second # Delivering work instruction block (a tighter variant) is added telling Claude to: interpret ambiguity as a careful colleague would; briefly state concerns then proceed as asked; finish the whole task not just the easy parts; stop short of actions clearly beyond what the request implies
All four use the shared helper gYt(envVar, flagName) which returns true if the env var is set, managed settings has the flag true, or Ke() returns true for the flag name

Evidence

Accessor functions and system prompt slots (search for "ekc()" near Fa_() for amber_astrolabe; "delivering_work_max" for bison_cairn; "overcorrection" for larch_cistern; "scope_fidelity" for alder_wicket)

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

Read from

Feature flag

tengu_bison_cairn Off in both readings

The flag server returned off for the account this site reads and for the anonymous baseline. A reading of off cannot rule out a rollout these two readings sit outside of.

This account: off · anonymous baseline: off · compiled default in v2.1.217: not a boolean we can read

These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.

Read once, for one account on one subscription tier, against v2.1.217. It isn't a statement about your account. What a flag value here can and cannot tell you

What has happened since

Flag reading moved 10 Sep 2026 The flag server now returns off for tengu_bison_cairn, read as this account. A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.

Flag reading moved 10 Sep 2026 The flag server now returns nothing at all, having stopped carrying it for tengu_bison_cairn, read as this account. A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.

See this across every release →

See this entry in the whole of v2.1.217 →

The build it came inv2.1.217

21 Jul 2026 · 27 entries, this one is #26.

Either side of it

v2.1.216Older
v2.1.218Newer

Open the release →

Names in this entry

flagtengu_amber_astrolabe
Default false. One of the switches that add the autonomy_append system prompt section ("You are operating autonomously...": proceed with reversible actions without asking, report findings without applying a fix when the user is only asking, finish the work before ending the turn). The section is added when tengu_amber_sextant is on (default true) and either the current model's settings turn it on or this check passes; the check passes when CLAUDE_CODE_AMBER_ASTROLABE is set, the client data key of this name is true, or this flag is on.
envCLAUDE_CODE_AMBER_ASTROLABE
Boolean (1, true, yes or on). When true, it forces the amber_astrolabe model capability on, which adds the system prompt section that opens You are operating autonomously. and tells the model to carry on with reversible actions without asking. The section still needs the tengu_amber_sextant flag (default true). When unset or false, the capability comes from the model's capability data, and the same section is also added for models with the fable_5_mitigations capability (on by default for claude-mythos-5).
flagtengu_bison_cairn
Default false. When on, the system prompt includes the # Delivering work section (guidance on doing the requested scope, handling uncertainty mid-task and proceeding after the user reaffirms a request). The section is also included when the cached client data has tengu_bison_cairn: true, for models with the fable_5_1_prompt_bundle capability (except under certain entrypoints), and for models with the opus_5_prompt_bundle capability while tengu_fennel_godwit is off; CLAUDE_CODE_BISON_CAIRN set to a true or false value overrides all of these.
envCLAUDE_CODE_BISON_CAIRN
A new env var forces the delivering-work-max prompt section on instead of relying on heuristics.
flagtengu_larch_cistern
Default false. When on, the system prompt gets an overcorrection section headed # Corrections, telling the model to avoid unnecessary self-correction, to correct an earlier statement only when the error would change the user's code, conclusions or decisions, and not to treat a follow-up question as a sign it was wrong. The section is also added when CLAUDE_CODE_LARCH_CISTERN is set, when the Opus 5 prompt bundle applies (unless tengu_fennel_godwit is on), or when the client data key of this name is true.
envCLAUDE_CODE_LARCH_CISTERN
SDK control commands like get_context_usage no longer need a special environment variable to be recognized
flagtengu_alder_wicket
Default false. When on, the system prompt gets a # Delivering work section (section id scope_fidelity) telling the model to deliver what was asked at the intended scope, finish the whole task and say plainly what is missing. The section is also turned on by a truthy CLAUDE_CODE_ALDER_WICKET env var or a tengu_alder_wicket value of true in the cached client data.
envCLAUDE_CODE_ALDER_WICKET
Boolean. When true, the system prompt includes a # Delivering work section (Deliver what the user asked for, at the scope they intended.). The section is also included when the tengu_alder_wicket key in cached client data is true or the tengu_alder_wicket flag is on.

Is this right?