Conciseness

Mostly efficient operational content, but the same epistemic principles (evidence before narrative, hedging, naming gaps, no intent from code shape) are stated three times across "Mindset", "Operating Posture", and "Core Epistemics" before the pointer to references/epistemics.md, and the opening blockquote's tokensave/MCP-mapping guidance is repeated in Step 2 and the Discovery subsection. Not anchor 2 because the majority of the 378 lines — tool commands, roster, skip rules, output format, failure modes — is dense non-redundant guidance, not padding.

Actionability

Fully executable, copy-paste-ready guidance: exact tokensave tool signatures (tokensave_blame/log/diff/changelog with parameters), concrete git fallbacks (git blame -L <start>,<end>, git log --follow -p), the full gh pr view --json field list, subagent config (subagent_type, model guidance), and a worked example of the Sources Consulted output lines.

Workflow Clarity

Steps 1–5 (Understand → Code Anchor → Spawn Investigators → Synthesize → Present) are clearly sequenced, with an explicit validation checkpoint in the Verification section ("spot-check at least one citation... Run this the same way you'd run a test before shipping") and feedback loops via the Troubleshooting and Common Failure Modes sections covering contradiction handling, empty results, and misapplied skips.

Progressive Disclosure

The body holds orchestration while details are split one level deep into verified, well-signaled files: epistemics.md (confidence framework), investigator-prompt.md and synthesizer-prompt.md (templates), and a per-category playbook index (source-playbook.md → sources/*.md, which SKILL.md also references directly), with a Reference Files section mapping each file to its purpose. All referenced paths exist in the bundle.

Specificity

The description lists multiple specific concrete actions covering the full pipeline: "Discovers available MCPs", "queries each evidence category (source control, issue tracker, long-form docs, real-time chat, infrastructure observability, error tracking, product analytics warehouse) in parallel", and "returns a cited read on decisions and tradeoffs" — the enumerated categories make the capability set comprehensive rather than generic.

Completeness

Both what and when are explicit: the what is "Discovers available MCPs and queries each evidence category... returns a cited read on decisions and tradeoffs", and the when is given twice — "Use for..." and the dedicated "Triggers:" clause with concrete trigger phrases, plus negative triggers for companion skills.

Trigger Term Quality

The trigger list — "why does X work this way, why did we pick Y over Z, what motivated this, why does this still exist, history of X" — plus "design rationale, regressions, postmortems, or data-backed thresholds" comprehensively covers the natural phrases a user would say, including synonyms. No file extensions apply to this skill's domain.

Distinctiveness Conflict Risk

Clear niche (design-rationale archaeology) with minimal conflict risk: the explicit negative triggers — "DO NOT use for explaining current runtime behavior (use how instead), or for proving a change is safe to ship (use blast-radius instead)" — proactively disambiguate it from the two most confusable sibling skills.