2.1.6 Test (2.1.6)
| # | Feature | Status | Key Finding |
Claude Code 2.1.6 is a UX polish release that introduced nested skills discovery and percentage-based context-window fields, alongside incremental improvements to /config, /stats, and /doctor.
Epistemic status: changelog-derived and partially code-reviewed; no runtime verification was completed. Two features were confirmed by code review; four interactive features were queued for manual testing but not executed during this study.
What the test examined
Testing was conducted on 2026-01-16 against version 2.1.9 (which carried the 2.1.6 feature set). Six features were targeted. The session could not runtime-verify any of them: four required interactive input (/config search filtering, /stats date-range cycling via r, /doctor update-channel display, Ctrl+G error surfacing), and the remaining two depended on context that the session could not replicate. Both of the latter were assessed by code review only.
The nested skills test attempted to verify automatic discovery by creating the structure:
2.1.6/tests/01-nested-skills/
└── .claude/
└── skills/
└── nested-test/
└── SKILL.md
Discovery could not be confirmed because skill resolution depends on the working-directory context of a live session; a fresh claude invocation inside that subdirectory would be required. Status: CODE REVIEW, not runtime-verified.
What the sources say about the two notable features
Nested skills discovery. Skills placed in a subdirectory's .claude/skills/ are now discovered automatically when working with files in that subtree. Previously, skills had to live at the project root. The architectural implication for monorepos is that each package can carry its own skill set without polluting root-level configuration.
"nested skills + hot-reload + agent field = legitimate monorepo support. each package can have its own skills in
package/.claude/skills/" -- OPUS-THOUGHTS.md
Status line percentage fields. Two new template variables were added to statusLineTemplate configuration:
{{context_window.used_percentage}}-- integer 0-100 representing tokens consumed{{context_window.remaining_percentage}}-- inverse
These complement the existing context_window.total_tokens field. Example usage confirmed by code review:
{ "statusLineTemplate": "{{context_window.used_percentage}}% used | {{context_window.remaining_percentage}}% free" }
Why it matters
Nested skills discovery changes the organizational model for large projects: skill authorship can be delegated per subtree rather than centralized. The percentage fields address a usability gap -- raw token counts require mental arithmetic against a model-specific limit, while percentages are immediately interpretable.
A security fix also shipped in this release: shell line-continuation characters could previously bypass command blocking; 2.1.6 closes that path. This was noted in the changelog but not independently verified.
Caveats
All findings here are changelog-derived or code-review-derived. No interactive feature was runtime-confirmed. The nested skills behavior in particular requires a session started inside the nested directory; the test scaffolding exists but the live verification step was deferred.
- file2.1.6/SUMMARY.md
release_63effe42e9e083aeb11b57dfed255199b87705613b1e2fd064d57fa75a6b679d2856ceafad6b1daa8f982493871b6dd716fb9ed011e1247577aacf5827bf44b60216796a5dd250f9dbdb8ded516a8405c9a73a8ae29d1004304295861baad4aca9b35dc5faf77047c4faa61e4db610eSigned with an ed25519 key held off the repo. Anyone can verify against the published public key; nobody without the secret key can forge it. Click verify: it recomputes the signature in your browser. The signature proves integrity and authorship of this exact content — not a third-party timestamp or that the underlying claim is objectively true. signedAt is when the @f3/attest pipeline ran, not when the work happened; the evidence refs carry the source dates.
- introduces Nested Skills Discovery Primitive
- introduces Status Line Percentage Fields Primitive