diff --git a/specs/superpowers-analysis.md b/specs/superpowers-analysis.md
deleted file mode 100644
index c22229f30..000000000
--- a/specs/superpowers-analysis.md
+++ /dev/null
@@ -1,149 +0,0 @@
-# Spec: Superpowers Pattern Analysis
-
-Extracted from `docs/superpowers/plans/` (now deleted) and external
-research. This spec documents what superpowers is, how it overlaps
-with ctx, and which patterns are worth adopting natively.
-
----
-
-## What is Superpowers?
-
-[obra/superpowers](https://github.com/obra/superpowers) is an
-open-source Claude Code plugin by Jesse Vincent. It installs as a
-skill pack via the Anthropic plugin marketplace and enforces a rigid
-workflow: brainstorm, plan, execute (with subagents), review.
-
-**Distribution:** Claude Code plugin marketplace (`obra/superpowers`).
-Also ported to Google Antigravity IDE, OpenCode, GitHub Copilot CLI,
-Gemini CLI.
-
-**Mechanism:** Each skill is a `SKILL.md` file with YAML frontmatter
-(metadata, triggers, tool permissions) and Markdown instructions.
-Skills auto-activate based on context. The `using-superpowers`
-meta-skill enforces that if any skill applies, the agent must use it.
-
----
-
-## Skill Overlap
-
-| Capability | ctx | Superpowers |
-|---|---|---|
-| Design-before-code | `/ctx-brainstorm` | `brainstorms` |
-| Plan writing | `/ctx-spec` | `writing-plans` |
-| Plan execution | `/ctx-implement` | `executing-plans` |
-| Subagent dispatch | `/ctx-worktree` | `subagent-driven-development` |
-| Git worktrees | `/ctx-worktree` | `using-git-worktrees` |
-| Code review | `/ctx-code-review` | `requesting-code-review` |
-| TDD workflow | `/ctx-refactor` | `test-driven-development` |
-| Debugging | `gitnexus-debugging` | `systematic-debugging` |
-| Skill creation | `/ctx-skill-creator` | `writing-skills` |
-
-**Key difference:** Superpowers is a pure instruction layer (no
-runtime, no persistence, no CLI). ctx is a full system with CLI,
-state management, persistence, and skills. Superpowers tells the
-agent *how to work*; ctx gives the agent *tools and memory*.
-
----
-
-## Patterns Worth Adopting
-
-### 1. Micro-task decomposition (from `writing-plans`)
-
-Superpowers enforces 2-5 minute tasks with exact file paths and
-verification steps. ctx's `/ctx-implement` already does stepwise
-execution, but the explicit time-boxing and mandatory verification
-step per task is worth codifying.
-
-**Action:** Consider adding a verification checkpoint to each step
-in `/ctx-implement` plans. Currently steps are executed sequentially
-but verification is ad-hoc.
-
-### 2. Subagent isolation (from `subagent-driven-development`)
-
-Each task gets a fresh subagent with no prior context bleed. The
-subagent receives only the task description, relevant file paths,
-and the plan. This prevents context pollution across tasks.
-
-**Action:** `/ctx-worktree` already uses git worktrees for isolation.
-The pattern of giving each subagent a minimal context packet (not
-the full conversation) is the valuable idea. Could improve agent
-dispatch prompts.
-
-### 3. Two-stage review (from `requesting-code-review`)
-
-After implementation, two review passes: (1) spec compliance — does
-it match the plan? (2) code quality — is it well-written? Separating
-these prevents quality concerns from masking spec violations.
-
-**Action:** `/ctx-code-review` currently does a single-pass review.
-Consider structuring it as two explicit passes.
-
-### 4. Mandatory skill activation (from `using-superpowers`)
-
-A meta-skill that enforces: if a skill applies, the agent must use
-it. Prevents the agent from doing ad-hoc work when a structured
-skill exists.
-
-**Action:** ctx already has skill trigger detection in skill
-descriptions. The gap is enforcement — agents can ignore triggers.
-The AGENT_PLAYBOOK already says "check available skills" but doesn't
-mandate usage.
-
-### 5. Observe-failure-first TDD (from `test-driven-development`)
-
-Superpowers enforces that the agent must *observe* the test failing
-before writing implementation. It will delete code written before
-tests. This is stricter than ctx's test-first guidance.
-
-**Action:** `/ctx-refactor` already says test-first. The
-observe-failure step is the useful addition — ensure the red phase
-is verified, not just written.
-
----
-
-## Patterns to Avoid
-
-### Token-hungry brainstorming
-
-Superpowers brainstorms spawn parallel agents for exploration,
-consuming large token budgets. ctx's `/ctx-brainstorm` is
-conversation-based and cheaper. The parallel exploration approach
-is expensive without proportional value for most tasks.
-
-### Rigid workflow enforcement
-
-Superpowers mandates brainstorm-before-any-code, even for trivial
-fixes. ctx should remain flexible — structured workflows for complex
-work, direct execution for simple changes.
-
-### Competing meta-skills
-
-Installing superpowers alongside ctx creates conflicting skill
-triggers (both have brainstorm, review, worktree skills). The
-systems are not designed to coexist.
-
----
-
-## How Superpowers Arrived in This Repo
-
-Murat Parlakisik's feature branch merge (`1d05145e`, 2026-04-03)
-included `docs/superpowers/plans/2026-03-31-commit-context-tracing.md`
-— an 88KB implementation plan generated by the superpowers
-`writing-plans` skill. The plan assumed superpowers was the execution
-engine (header: "REQUIRED SUB-SKILL: Use
-superpowers:subagent-driven-development").
-
-The feature it planned (commit context tracing) already had a proper
-ctx-native spec at `specs/commit-context-tracing.md`. The superpowers
-plan was redundant and introduced an external dependency.
-
-**Resolution:** Superpowers plan file deleted. This analysis spec
-preserves the useful patterns. The commit-context-tracing feature
-proceeds from its existing ctx-native spec.
-
----
-
-## Status
-
-Analysis complete. No action items created — patterns noted above
-are candidates for future skill improvements, not immediate work.