chore: release v0.11.5

[vyncint]

Version 0.11.5 in the workspace and the CLI's library dependency,
report@v0.11.5 in the three pinned files, SKILL.md written against
0.11.5, the bug report form's placeholder, and the CHANGELOG section
moved with its link definitions. tests/compat/0.11.5/ is 0.11.4's twelve
files byte for byte: the library's rendering did not change.

Release notes (extract-changelog.sh 0.11.5):

Changed

• The documentation was reviewed against the code, and each correction was
  checked by reading the code or running it:
  • README.md and the skill recommend cargo install termlens-cli --locked, the dependency set a release is tested with (docs: recommend cargo install termlens-cli --locked #552), and
    AGENTS.md's layout lists the termlens-cli crate (docs: list termlens-cli in AGENTS.md's layout #553).
  • The skill's rule 9 gives the true reason to return
    termlens::Result: a failing test prints the error's Debug form,
    screen included, as unwrap() would. §6 now reads that form as well
    as Display, and its unanswered-probe row links to
    docs/LIMITATIONS.md (docs(skill): rule 9 gives the true reason, and §6 reads the Debug form too #559). Rule 10, §6 and the mask row no longer
    say that format!("{:?}") equals Display, that every error carries
    a screen, or that mask_matching matches across a soft wrap (docs: a review pass, fifteen claims the code contradicts #562).
  • The rustdoc gains runnable examples for Screen::mask_matching (docs: add a runnable example to Screen::mask_matching #551)
    and ScreenDiff::changed_rows and style_changes (docs: ScreenDiff's changed_rows and style_changes get runnable examples #560). On docs.rs,
    Unsupported no longer shows Clipboard's summary, and Clipboard
    has one of its own (docs: a review pass, fifteen claims the code contradicts #562).
  • docs/DESIGN.md, CONTRIBUTING.md, docs/RELEASING.md,
    docs/LIMITATIONS.md and SECURITY.md now match the code. They cover
    the two translated character sets and DECSC saving the charset state,
    the SHELL=/bin/sh pin under env_clear, the tools/ directory, and
    the release checklist's two ungated copies of the version (docs: DESIGN, CONTRIBUTING and RELEASING catch up with the code #558). They
    also cover the unbounded writer queue, which CSI … t reports are
    answered, and Dependabot's carve-out from the bot-identity rule (docs: a review pass, fifteen claims the code contradicts #562).

Fixed

• A relative TerminalBuilder::current_dir, and so termlens inspect --cwd ./dir, runs a relative program path such as ./myapp inside that
  directory. The directory was applied twice — the program was looked up
  as dir/./myapp from inside dir — and the child died of a Rust runtime
  abort it could not report (fatal runtime error: assertion failed: output.write(&bytes).is_ok()), which inspect showed as exited: killed by signal: Aborted. A relative directory is now taken from the test
  process's own, as std::process::Command does. An absolute directory,
  or a bare program name found on PATH, was never affected.
• termlens inspect --help and examples/inspect.rs --help describe the
  wait and the environment as they are: the silence window starts only
  once the program has drawn something; the child also gets
  TERM=xterm-256color and SHELL=/bin/sh; and --inherit-env still
  replaces the caller's TERM unless --env TERM=… sets it. --cwd's
  argument is called DIR in the help and in its diagnostic, beside a
  paragraph about the PATH variable, and the help says a relative program
  path is found inside it. The CLI README's --cwd example ran a program
  from the wrong directory and is corrected.

Local checks on this tree: tools/preflight.sh 35 ok, 1 skipped (zizmor: pipx not installed here; CI runs it), 0 failed, semver patch-compatible with published 0.11.4. CLI behaviour diff against published termlens-cli 0.11.4: 2,854 cases, 6 differ, all in the Fixed section (three inspect help texts, two relative --cwd runs that 0.11.4 aborts, and the --cwd diagnostic).