Skip to content

nanaagyei/noetherkin / skills/user-agent

user-agent

Verified

Provide bounded simulated-user feedback on explicitly shared Noetherkin product behavior. Use for user-perspective exercises, not real-demand claims or learner evaluation access.

FileKindSize
SKILL.mdskill md1.1 KB
assets/proposal.mdmarkdown2.0 KB
references/FOUNDATION_V1.mdmarkdown5.4 KB
references/OPEN_QUESTIONS.mdmarkdown3.5 KB
references/PROJECT_CHARTER.mdmarkdown6.4 KB
references/apprenticeship-config.schema.jsonconfig3.0 KB
references/architectural-invariants.mdmarkdown2.3 KB
references/assessment.schema.jsonconfig5.0 KB
references/assistance-model.mdmarkdown4.0 KB
references/attention-advisory.schema.jsonconfig3.8 KB
references/bundle.jsonconfig12.0 KB
references/cli.mdmarkdown18.6 KB
references/competencies.yamlconfig37.7 KB
references/competency-model.mdmarkdown5.0 KB
references/competency-state.schema.jsonconfig3.4 KB
references/conformance.mdmarkdown22.0 KB
references/contract-README.mdmarkdown7.0 KB
references/contract-production-readiness.mdmarkdown1.5 KB
references/contract-task-assignment.mdmarkdown3.2 KB
references/contract-user-agent.mdmarkdown1.6 KB
references/current-project.schema.jsonconfig2.3 KB
references/current-track.schema.jsonconfig1.5 KB
references/evidence-model.mdmarkdown4.1 KB
references/evidence.schema.jsonconfig10.3 KB
references/forge.schema.jsonconfig9.5 KB
references/index.mdmarkdown3.6 KB
references/judgment-validity.mdmarkdown6.6 KB
references/learner-profile.schema.jsonconfig1.5 KB
references/leveling-framework.mdmarkdown3.4 KB
references/levels.yamlconfig8.0 KB
references/overview.mdmarkdown2.9 KB
references/permissions-model.mdmarkdown6.7 KB
references/project.schema.jsonconfig9.4 KB
references/proposals.mdmarkdown4.2 KB
references/review-model.mdmarkdown7.2 KB
references/review.schema.jsonconfig13.0 KB
references/runtime.mdmarkdown7.8 KB
references/selection-model.mdmarkdown6.1 KB
references/simulation-lifecycle.mdmarkdown8.0 KB
references/state-model.mdmarkdown13.7 KB
references/task.schema.jsonconfig13.8 KB
references/track.schema.jsonconfig1.3 KB

references/review-model.md

View on GitHub
# Reviews and promotion authority

All review records are immutable, evidence-linked judgments. Code and task reviews carry `change` identifying the reviewed deliverable at a fixed revision, including documentation deliverables. Task review publication must follow the current validation checks. An outcome applies to a named subject and a bounded interval. The period start must not follow its end or publication time. A later review supersedes an earlier same-kind/same-subject review when correcting its judgment; task rework ordinarily creates a fresh review for the new revision. For code/task gates, current review means latest unsuperseded review for the task and current artifact revision; equal-time conflicting heads are invalid. Performance packets use explicit learner/period/scope identity; corrections to that same packet must supersede its prior head. Promotion decisions use the adjacent-edge and reconciliation rules, never a generic latest-review selector.

| Kind | Inputs / subject | Author and output | Read / write boundary |
| --- | --- | --- | --- |
| code | Change artifact at fixed revision, task criteria, tests, evidence; task ID | Peer engineer: approve, changes-requested, insufficient-evidence | Read source/task/evidence; append code review only |
| task | Execution history, validation, assistance, current code review if required; task ID | Team lead: accepted, rework, insufficient-evidence | Read project work; append task review and govern task gates |
| performance | Evidence/assessments/reviews over a stated period; learner ID | Manager: continue, adjust-scope, recommend-promotion-review, insufficient-evidence | Read formal records; append performance review, no level writes |
| promotion | Current level chain, performance recommendation, technical assessment, rubric and evidence; learner ID | Promotion reviewer: PROMOTE, NOT YET, INSUFFICIENT EVIDENCE | Read formal records; append promotion review only |

Code approval neither merges a change nor grants task completion. A manager adjusts future task scope by recommendation to team lead, not by rewriting active acceptance criteria. Team lead provides technical input but cannot grant promotion alone.

## Promotion procedure

1. Learner and team lead agree the evaluation_scope before the review period. Team lead publishes a longitudinal readiness assessment covering exactly the core and agreed specialization competencies. Promotion `readiness_assessment_id` designates it, `assessment_ids` includes it, and `technical_reviewer_id` names its author.
2. Manager publishes a performance review recommending promotion review using that readiness assessment and scope. The caller explicitly selects this unsuperseded supported recommendation for the learner and exact review period and records its ID; an absent or conflicting recommendation blocks the procedure. Later overlapping reviews do not silently replace this dependency.
3. Learner explicitly authorizes that review and its potential adjacent-level decision. Store the authorization artifact and timestamp; this is process authorization, not proof of readiness.
4. A registered promotion reviewer with a principal ID different from the technical reviewer, manager recommender and learner examines supporting and contrary evidence. A distinct invocation with a distinct principal and independently recorded judgment is required; a second model vendor is not required. This separation is procedural, not a claim of bias-free independence.
5. Record `from_level` equal to the current derived effective level (except the explicit reconciliation procedure in judgment validity) and `target_level` exactly one step higher, with the previous accepted promotion pointer. Include findings for each of the thirteen rubric dimensions exactly once, evidence references and quality analysis for repetition, recency, diversity, scope, independence and contrary evidence.
6. PROMOTE requires each dimension to meet the target and relevant core/specialization coverage, with repeated independent demonstrations in materially different contexts. No numeric task quota substitutes for this judgment. Evidence must be recent enough for the claimed capability; the reviewer explains why rather than applying an arbitrary expiry window. Unresolved material contrary evidence precludes PROMOTE.
7. NOT YET means adequate evidence shows a next-level requirement is not met. INSUFFICIENT EVIDENCE means missing, unverifiable or contextually narrow evidence prevents a decision. Both leave effective level unchanged and name the next observation needed.
8. Deterministic checks validate identity, permissions, chain, evidence references and completeness before publication. Then rebuild derived competencies. Structural validity does not establish sound judgment.

Review evidence may precede the review period where explicitly needed for longitudinal comparison; identify that in findings. Observations and cited records cannot be from after review publication. Strong same-artifact repetitions or duplicate reviews are not independent demonstrations. A reviewer lacking required independent principal/context returns a blocked procedure, not a synthetic second reviewer. There is no promotion skill among the first eight; the protocol is reserved for a future dedicated workflow or a manually conducted conforming review.

## Frozen evaluation packet

Performance reviews carry `evaluation_scope`: exact core and agreed specialization competency sets, competency/level catalog versions, and an immutable scope_agreement artifact recording learner and team-lead agreement before the period starts. Core set equals all required_core IDs in the pinned catalog; specialty IDs are valid noncore IDs. Changing specialties requires a rationale and a fresh agreed packet/performance review, never mutation of an existing packet. Scope agreement is inspected, not inferred from the current profile.

A promotion carries an identical evaluation_scope, recommendation_review_id and readiness_assessment_id. The recommendation must be kind performance, same learner and exact period, outcome recommend-promotion-review; it must include the readiness ID in its assessment_ids. Readiness must be a supported longitudinal team-lead assessment for this learner, with findings covering exactly the packet's core and specialty union. Readiness is published no later than the recommendation, and recommendation no later than authorization and promotion. Authorization artifact binds learner, level edge, previous promotion ID, recommendation ID, readiness ID and evaluation scope. Evidence may predate the period with explicit rationale; none may postdate publication.

Technical author, recommending manager and promotion author must be pairwise distinct and each different from the learner principal, with the required role at their own publication. Subsequent profile edits, ordinary later reviews and retirement do not alter historical packet selection. Explicit correction, lost support or relevant contrary evidence propagates under [judgment validity](judgment-validity.md). PROMOTE requires all thirteen unique dimension findings to meet, supported findings for every required competency at target scope, and addressed contrary evidence. No runtime chooses historical dependencies by today's latest timestamp.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall