Implement 0086 PromptManager default cache TTL

[chris-colinsky]

Phase A of the v0.17.0 cycle (proposal 0086, prompt-management §6, spec v0.79.0).

PromptManager construction accepts an optional default_cache_ttl_seconds, applied to any fetch or get that omits a per-call cache_ttl_seconds. Resolution follows a precedence chain:

1. an explicit per-call value (including 0 force-fresh) wins;
2. otherwise the manager default;
3. otherwise None, so the backend's own caching governs.

An omitted or explicit-None per-call value both select the default, so resolution does not depend on argument presence. A negative default is rejected at construction; the per-call negative rejection is unchanged. get delegates to fetch and inherits the chain, and the bundled caching backends already honor a resolved cache_ttl_seconds (0072), so no backend changed.

Conformance

Un-defers fixture 036 at both the runtime and parse layers. Two harness additions carry the fixture's shape: FixtureManagerSpec gains default_cache_ttl_seconds and an optional backends (036's manager: block omits it and defaults to all declared backends), and a new ManagerTarget models the {manager: true} dict target form (Literal[True], so {manager: false} is rejected).

Tests / docs

Seven unit tests pin each precedence branch at the backend boundary (default-on-omit, explicit-None-selects-default, per-call-0 override, no-default-forwards-None, zero default, negative reject, and get() inheritance), plus fixture 036 end-to-end (default applied, clock aging across the boundary, per-call 0 override). Docs updated in concepts/prompts.md; new ## [0.17.0] changelog section.

Reviewed with /review and /adversarial-review; all actionable findings folded in. Full suite green.

[Copilot]

Pull request overview

Implements proposal 0086’s service-wide default cache TTL for PromptManager, aligning runtime behavior and conformance harness support with spec v0.79.0.

Changes:

• Add default_cache_ttl_seconds to PromptManager and apply precedence resolution when forwarding cache_ttl_seconds to backends.
• Un-defer and support prompt-management fixture 036 by extending the conformance harness models and runner target shapes.
• Add unit + conformance coverage and document the new default TTL behavior; mark proposal 0086 implemented in conformance.toml.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/unit/test_prompts.py	Adds unit tests that assert TTL precedence at the backend boundary.
tests/conformance/test_prompt_management.py	Updates fixture runner to support {manager: true} targets and manager defaults; un-defers fixture 036.
tests/conformance/test_fixture_parsing.py	Removes parsing-layer deferral for fixture 036.
tests/conformance/harness/prompt_management.py	Extends fixture schema with optional backends, default_cache_ttl_seconds, and ManagerTarget.
src/openarmature/prompts/manager.py	Adds default_cache_ttl_seconds to PromptManager and resolves TTL precedence in fetch (affecting get).
docs/concepts/prompts.md	Documents service-wide default TTL behavior and precedence.
conformance.toml	Marks proposal 0086 as implemented since 0.17.0 with explanatory note.
CHANGELOG.md	Adds 0.17.0 entry describing the new default TTL feature.