tacticaldoll
UserPortable, single-purpose skills for coding agents that orient, map, plan, review, and reason over your code — read / plan / report, never editing behind your back.
Categories
Indexed Skills (21)
assess-dependency
Use when an agent needs to decide whether to adopt a structural dependency into a project; checks local pain, architectural fit, feasibility, and governance, then returns adopt, adopt narrowly, defer, or decline rather than planning implementation.
assess-knowledge
Use when an agent needs to assess a conversation for knowledge worth extracting; identifies topics, maturity, and internalize/externalize tendency without creating artifacts or recommending next steps.
design-boundaries
Use when an agent needs to design component boundaries for a feature or system before the code exists; partitions responsibilities into single-purpose components with minimal interfaces and an acyclic dependency direction, then pressure-tests the boundaries rather than editing code, splitting an existing tangled unit, or extracting a separate repository.
explore-intent
Use when an agent needs to explore intent, requirements, and direction through dialogue before any building; holds a no-stakes, codebase-grounded thinking-partner stance that surfaces options and tradeoffs one question at a time and resists premature convergence, rather than implementing, designing component structure, or planning the work.
handle-feedback
Use when an agent needs to handle code-review or technical feedback with rigor rather than performance; verifies each point against the codebase before accepting, restates and pushes back with reasoning where warranted, and takes feedback one item at a time rather than performing agreement, implementing blindly, or applying the changes itself.
harden-skill
Use when an agent needs to harden a skill's instructions so they reliably change agent behavior; diagnoses each intended behavior's baseline failure, matches it to the right instruction form, hardens discipline rules against rationalization, and specifies a micro-test protocol rather than editing the skill or running the tests itself.
map-codebase
Use when an agent needs to understand how an unfamiliar codebase or subsystem actually works before changing it; maps entry points, control and data flow, key abstractions, and external effects to the depth the task needs, and reports what was not traced rather than reviewing quality or editing code.
orient-repo
Use when an agent needs to orient itself in an unfamiliar repository before acting; discovers governance, conventions, enforcement, and non-goals into a compact working brief rather than editing files or relying on one host's auto-loaded context.
plan-implementation
Use when an agent needs to turn a goal into an ordered, verifiable implementation plan before coding; slices the work into the smallest independently verifiable increments, sequences them by dependency and risk, and names each step's acceptance check rather than writing code, designing component structure, or planning a behavior-preserving split.
plan-migration
Use when an agent needs to plan a safe data migration or database schema change; sequences the change into backward-compatible expand/backfill/contract phases with rollback and verification steps rather than executing SQL scripts or migrating data.
plan-repo-extract
Use when an agent needs to assess whether a cohesive part of a repository should become an independent repository; evaluates extractability and plans dependency, API, history, and migration work rather than executing the split.
plan-split
Use when an agent needs to plan splitting a large or tangled code unit inside the same codebase; finds natural responsibility boundaries, dependency risks, and reviewable extraction steps rather than editing code or executing the refactor.
plan-testing
Use when an agent needs to design a comprehensive test strategy for a feature or refactor; identifies unit boundaries, mocks, and edge cases rather than writing the actual test code or executing test suites.
resolve-deadlock
Use when an agent needs to resolve conflicting requirements or a governance deadlock that is already blocking progress; facilitates a structured negotiation to find a pragmatic third path rather than halting execution or unilaterally rewriting project governance.
save-knowledge
Use when an agent needs to persist mature conversation knowledge into durable project, agent, or team sources; chooses the narrowest destination and drafts the edit for downstream execution rather than internalizing it for personal understanding.
scope-new-skill
Use when an agent needs to explore whether a vague workflow or knowledge process should become a skill; classifies skill type, outputs, resources, risks, and readiness rather than creating or editing skill files.
static-review
Use when an agent needs to review local or pasted code with sequential quality gates; resolves files, diffs, commits, branches, or snippets into a static review scope and reports findings without modifying code or calling remote services.
write-learning-report
Use when an agent needs to turn mature conversation content into a self-contained learning report; selects a fitting report structure and preserves source fidelity for personal understanding rather than durable project knowledge.
assess-threats
Use when an agent needs to identify trust boundaries and architectural threat vectors in a system design or a change's attack surface; maps the boundaries, applies STRIDE, and rates threats by risk rather than executing exploits or patching code.
diagnose-issue
Use when an agent needs to find a bug's root cause from a report or stack trace; statically traces control and data flow, identifies the logical fault type, and produces a confidence-rated root-cause hypothesis with a verification step, rather than editing code or attempting a fix.
replace-me
Use when an agent needs to perform a specific workflow. Replace this with clear trigger contexts and the capability this skill provides.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.