Sequence & representation
Protein Language Model
Learn numerical representations of residues and proteins from sequences.
Representations and model scores are not experimental function.
METHODS / WORK RECORDS
Keep the record. Carry the work forward.
Original method notes and work snapshots, with their evidence, dates and limits intact. These records do not imply newly completed work.
WUTONG / A NOTEBOOK OF METHODS
The Wutong skill tree
Methods in the leaves. Questions meeting between branches.
THE NOTEBOOK IN THE LEAVES
Start with a leaf.
Choose a leaf to explore its method, tools and connections.
Choose a method
Structure & interaction
Structure prediction
Infer candidate three-dimensional structures from model inputs.
Predicted structures and pLDDT do not validate binding.
UPSTREAM IMPLEMENTATIONS
AlphaFold 3Structure prediction entry ↗Public source entry points · method notes, not claims of execution or mastery.
MY PUBLIC PRACTICE
A Day, Folded · prediction archive2026-10-13 ↗An existing public prediction record. Explore its structure and confidence; it does not establish binding or experimental efficacy.
View public structure data ↗PUBLIC EVIDENCE & GUIDANCE
Structure & interaction
Biomolecular docking
Model possible spatial arrangements of interacting molecules.
Docking poses are hypotheses to test.
PUBLIC EVIDENCE & GUIDANCE
Design & optimization
Inverse folding
Design amino acid sequences conditioned on a given backbone.
Model compatibility scores do not establish affinity.
UPSTREAM IMPLEMENTATIONS
ProteinMPNNSequence design entry ↗Public source entry points · method notes, not claims of execution or mastery.
PUBLIC EVIDENCE & GUIDANCE
Design & optimization
Generative backbone design
Propose candidate backbones using generative models.
Sequence design, evaluation and experiments remain necessary.
UPSTREAM IMPLEMENTATIONS
RFdiffusionBackbone sampling entry ↗Public source entry points · method notes, not claims of execution or mastery.
PUBLIC EVIDENCE & GUIDANCE
Design & optimization
Partial diffusion
Noise and denoise an existing backbone to sample nearby designs.
Diversification does not guarantee improvement.
UPSTREAM IMPLEMENTATIONS
RFdiffusionBackbone sampling entry ↗Public source entry points · method notes, not claims of execution or mastery.
PUBLIC EVIDENCE & GUIDANCE
Design & optimization
Interface redesign
Propose interface sequence variants while keeping selected structure or positions fixed.
Constrained design still needs independent evaluation and binding assays.
UPSTREAM IMPLEMENTATIONS
ProteinMPNNSequence design entry ↗Public source entry points · method notes, not claims of execution or mastery.
PUBLIC EVIDENCE & GUIDANCE
Design & optimization
Rational design
Propose reasoned sequence changes using structure and position constraints; this entry covers constrained sequence design.
A design rationale is not experimental evidence.
UPSTREAM IMPLEMENTATIONS
ProteinMPNNSequence design entry ↗Public source entry points · method notes, not claims of execution or mastery.
PUBLIC EVIDENCE & GUIDANCE
Physics & energy
Computational mutation scanning
Score sequence variants with a model to nominate candidates for testing.
Computational scores and experimental mutation effects are recorded separately.
UPSTREAM IMPLEMENTATIONS
ESM-2Variant scoring example ↗Public source entry points · method notes, not claims of execution or mastery.
PUBLIC EVIDENCE & GUIDANCE
Physics & energy
Molecular dynamics
Simulate molecular motion under a specified physical model.
Trajectories depend on sampling, initial conditions and force fields.
UPSTREAM IMPLEMENTATIONS
OpenMMDynamics example ↗Public source entry points · method notes, not claims of execution or mastery.
PUBLIC EVIDENCE & GUIDANCE
Selection & experiment
Experimental mutation scanning
Measure mutation effects using variant libraries and experimental readouts.
Readouts depend on assay conditions and are not universal affinity measurements.
PUBLIC EVIDENCE & GUIDANCE
Selection & experiment
mRNA Display Analysis
Analyze sequence counts, enrichment and candidate families across selection rounds.
Enrichment also reflects amplification and selection biases; it is not affinity.
PUBLIC EVIDENCE & GUIDANCE
New methods earn new leaves. Implementation improvements and validation deepen existing notes. Growth follows evidence.
Full work records
VIRTUAL R&D / TASK ORGANIZATION SNAPSHOT
Work in progress.
Research goals, current stages and next steps.
Reading views
Research platform · Current engineering proposals and follow-upsMerged
CI worker and workflow maintenance
Evidence date:
Integrate changes intended to improve CI worker availability and simplify workflow configuration.
Constraints & next steps
- Goal
- Integrate changes intended to improve CI worker availability and simplify workflow configuration.
- Constraints
- A source merge was observed; deployment, sustained operation, test success and scientific effectiveness are not established.
- Next step / knowledge feedback
- Check deployment, regression and actual-use evidence separately within the applicable scope.
Research platform · Current engineering proposals and follow-upsPlanned
Declare a regression-check baseline dependency
Evidence date:
Document which code path consumes a baseline in automated checks.
Constraints & next steps
- Goal
- Document which code path consumes a baseline in automated checks.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsPlanned
Version benchmark definitions and views
Evidence date:
Propose versioned benchmark definitions and ways to inspect their records.
Constraints & next steps
- Goal
- Propose versioned benchmark definitions and ways to inspect their records.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsMerged
Separate batch and resident compute interfaces
Evidence date:
Reduce coupling between batch work and persistent-service interfaces while organizing older entry points.
Constraints & next steps
- Goal
- Reduce coupling between batch work and persistent-service interfaces while organizing older entry points.
- Constraints
- A source merge was observed; deployment, sustained operation, test success and scientific effectiveness are not established.
- Next step / knowledge feedback
- Check deployment, regression and actual-use evidence separately within the applicable scope.
Research platform · Current engineering proposals and follow-upsPlanned
Plan evaluation-record views and interfaces
Evidence date:
Propose inspection views and access interfaces for evaluation records.
Constraints & next steps
- Goal
- Propose inspection views and access interfaces for evaluation records.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsPlanned
Plan layered evaluation coverage
Evidence date:
Propose layers of checks for software and research workflows.
Constraints & next steps
- Goal
- Propose layers of checks for software and research workflows.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsMerged
Integrate a design-method adapter
Evidence date:
Connect an additional design method through a provider interface.
Constraints & next steps
- Goal
- Connect an additional design method through a provider interface.
- Constraints
- A source merge was observed; deployment, sustained operation, test success and scientific effectiveness are not established.
- Next step / knowledge feedback
- Check deployment, regression and actual-use evidence separately within the applicable scope.
Research platform · Current engineering proposals and follow-upsPlanned
Plan a bounded data-catalog feasibility check
Evidence date:
Define the scope and feasibility questions for a bounded data-catalog check.
Constraints & next steps
- Goal
- Define the scope and feasibility questions for a bounded data-catalog check.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsDeveloping
Review monitoring storage and recovery
Evidence date:
Review storage, backup and recovery changes for scheduled monitoring.
Constraints & next steps
- Goal
- Review storage, backup and recovery changes for scheduled monitoring.
- Constraints
- This is an open non-draft code proposal; developing denotes pending code review, not live execution or passing tests.
- Next step / knowledge feedback
- Reconcile code review and original check evidence before deciding on integration.
Research platform · Current engineering proposals and follow-upsDeveloping
Make a planning reference portable
Evidence date:
Change a planning reference to reduce dependence on a particular local location.
Constraints & next steps
- Goal
- Change a planning reference to reduce dependence on a particular local location.
- Constraints
- This is an open non-draft code proposal; developing denotes pending code review, not live execution or passing tests.
- Next step / knowledge feedback
- Reconcile code review and original check evidence before deciding on integration.
Research platform · Current engineering proposals and follow-upsPlanned
Plan durable request evidence and recovery
Evidence date:
Propose interfaces to preserve and recover request evidence before retrying.
Constraints & next steps
- Goal
- Propose interfaces to preserve and recover request evidence before retrying.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsPlanned
Plan recovery for expired service resources
Evidence date:
Propose controlled recovery when a service resource allocation expires.
Constraints & next steps
- Goal
- Propose controlled recovery when a service resource allocation expires.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsPlanned
Plan controlled continuation after publication
Evidence date:
Propose explicit stopping and continuation boundaries after publication.
Constraints & next steps
- Goal
- Propose explicit stopping and continuation boundaries after publication.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsPlanned
Plan evidence checks for predictive-method research
Evidence date:
Propose evidence checks for predictive-method research and its written account.
Constraints & next steps
- Goal
- Propose evidence checks for predictive-method research and its written account.
- Constraints
- This is an open draft code proposal, not merged; it does not establish that tests, models or services are running.
- Next step / knowledge feedback
- Review the proposal scope and retain actual evidence for later implementation checks.
Research platform · Current engineering proposals and follow-upsPlanned
Track workflow dependency provenance
Evidence date:
Record dependency-provenance questions remaining after workflow changes.
Constraints & next steps
- Goal
- Record dependency-provenance questions remaining after workflow changes.
- Constraints
- This is an open follow-up; its title supports a general scope, not evidence that work started, passed verification or was resolved.
- Next step / knowledge feedback
- Check the issue and subsequent evidence within the authorized scope before implementation decisions.
Research platform · Current engineering proposals and follow-upsPlanned
Clarify check-planning boundaries
Evidence date:
Clarify applicability, refusal explanations and coverage declarations in check planning.
Constraints & next steps
- Goal
- Clarify applicability, refusal explanations and coverage declarations in check planning.
- Constraints
- This is an open follow-up; its title supports a general scope, not evidence that work started, passed verification or was resolved.
- Next step / knowledge feedback
- Check the issue and subsequent evidence within the authorized scope before implementation decisions.
Research platform · Current engineering proposals and follow-upsPlanned
Plan method qualification across repeated runs
Evidence date:
Record qualification needs involving execution checks, adoption and repeated-run evidence.
Constraints & next steps
- Goal
- Record qualification needs involving execution checks, adoption and repeated-run evidence.
- Constraints
- This is an open follow-up; its title supports a general scope, not evidence that work started, passed verification or was resolved.
- Next step / knowledge feedback
- Check the issue and subsequent evidence within the authorized scope before implementation decisions.
Research platform · Current engineering proposals and follow-upsPlanned
Plan recovery of missing method assets
Evidence date:
Track recovery and regeneration questions for missing method assets.
Constraints & next steps
- Goal
- Track recovery and regeneration questions for missing method assets.
- Constraints
- This is an open follow-up; its title supports a general scope, not evidence that work started, passed verification or was resolved.
- Next step / knowledge feedback
- Check the issue and subsequent evidence within the authorized scope before implementation decisions.
Personal engineering · Current engineering proposals and follow-upsPlanned
Plan recovery after an interrupted installation
Evidence date:
Track safe recovery after an interrupted first installation.
Constraints & next steps
- Goal
- Track safe recovery after an interrupted first installation.
- Constraints
- This is an open follow-up; its title supports a general scope, not evidence that work started, passed verification or was resolved.
- Next step / knowledge feedback
- Check the issue and subsequent evidence within the authorized scope before implementation decisions.
Papers and methods. Research ↗ Capability catalog ↗
Retained historical snapshot
VIRTUAL R&D / TASK ORGANIZATION SNAPSHOT
Work in progress.
Research goals, current stages and next steps.
Historical records · awaiting an update.
Reading views
Personal engineering · Personal tool engineeringMerged
Interactive registry shape validation
Evidence date:
Allow an optional state omission in unrelated interactive records while strictly checking the selected record.
Constraints & next steps
- Goal
- Allow an optional state omission in unrelated interactive records while strictly checking the selected record.
- Constraints
- Code is merged; a missing field cannot establish task state or scientific completion.
- Next step / knowledge feedback
- Continue checking malformed shapes, duplicate identities and selected-record boundaries.
Personal engineering · Personal tool engineeringMerged
Identity checks before message sending
Evidence date:
Narrow native process identity checks and retain refusal when evidence is incomplete.
Constraints & next steps
- Goal
- Narrow native process identity checks and retain refusal when evidence is incomplete.
- Constraints
- Code is merged; current deployment, successful delivery and elimination of the observation-to-send race are not established.
- Next step / knowledge feedback
- Verify deployment, read-only checks and operational applicability separately.
Research platform · Research platform and methodsMerged
Abort records for unsubmitted tasks
Evidence date:
Record explicit aborts for unsubmitted tasks without fabricating execution handles or completion results.
Constraints & next steps
- Goal
- Record explicit aborts for unsubmitted tasks without fabricating execution handles or completion results.
- Constraints
- Abort-handling code is merged; engineering records do not establish that scientific work ran or was effective.
- Next step / knowledge feedback
- Continue checking abort, recovery and existing authorization boundaries.
Personal engineering · Personal tool engineeringMerged
Reminder reading and calendar coverage
Evidence date:
Reuse reminder read results and preserve coverage evidence and last-good records for incomplete calendar collection.
Constraints & next steps
- Goal
- Reuse reminder read results and preserve coverage evidence and last-good records for incomplete calendar collection.
- Constraints
- Code merge is confirmed; complete local reads establish neither upstream synchronization nor long-term stability or speedup.
- Next step / knowledge feedback
- Continue checking natural-run coverage and upstream synchronization differences.
Personal engineering · Personal tool engineeringMerged
Metadata pagination and coverage
Evidence date:
Support bounded pagination and resumption while keeping missing identities unknown.
Constraints & next steps
- Goal
- Support bounded pagination and resumption while keeping missing identities unknown.
- Constraints
- Pagination code is merged; missing identity cases still require handling.
- Next step / knowledge feedback
- Complete missing-identity handling and runtime validation.
Tool · Metadata pagination
Research platform · Research platform and methodsMerged
Capability-linked engineering checks
Evidence date:
Make developer guidance explicit about test maintenance, applicability and unexecuted coverage.
Constraints & next steps
- Goal
- Make developer guidance explicit about test maintenance, applicability and unexecuted coverage.
- Constraints
- This merge updates developer guidance; it establishes no new scheduling behavior or scientific validation.
- Next step / knowledge feedback
- Continue checking test selection and coverage gaps against each change.
Personal engineering · Personal tool engineeringMerged
Offline message intake
Evidence date:
Quarantine minimal message metadata with replay and conflict checks.
Constraints & next steps
- Goal
- Quarantine minimal message metadata with replay and conflict checks.
- Constraints
- Metadata intake code is merged; attachment archiving and continuing integration remain incomplete.
- Next step / knowledge feedback
- Complete archive and integration validation within the applicable scope.
Tool · Offline message intake
Research platform · Research platform and methodsMerged
Tool guidance and wiring maintenance
Evidence date:
Remove stale guidance and align tool entry points with maintenance checks.
Constraints & next steps
- Goal
- Remove stale guidance and align tool entry points with maintenance checks.
- Constraints
- Code and guidance are merged; runtime behavior still needs checks for each task.
- Next step / knowledge feedback
- Continue checking tools against their guidance.
Tool · Tool wiring checks
Research platform · Research platform and methodsMerged
Optimization method routing
Evidence date:
Align optimization entry points, configurable roles and usage guidance.
Constraints & next steps
- Goal
- Align optimization entry points, configurable roles and usage guidance.
- Constraints
- Guidance and entry points are aligned; no new model or molecular results are claimed.
- Next step / knowledge feedback
- Check routing and applicability through real tasks.
Skill · Peptide optimization methods
Research platform · Research platform and optimization workflowsMerged
Lead optimization integration
Evidence date:
Connect optimization methods through shared constraints, candidate provenance, feedback and recovery.
Constraints & next steps
- Goal
- Connect optimization methods through shared constraints, candidate provenance, feedback and recovery.
- Constraints
- The optimization interfaces are merged; historical case replay is incomplete and scientific effectiveness remains unverified.
- Next step / knowledge feedback
- Complete historical case recovery and replay before evaluating method effectiveness.
Algorithm · Sequence-gradient optimization: adjust a sequence representation using model feedback within declared constraints.Algorithm · Local diffusion: explore constrained local changes around an existing candidate.Algorithm · Mutation scanning: compare bounded single or combined changes.Algorithm · Backbone-conditioned MPNN: generate candidate sequences conditioned on a backbone.Algorithm · MPNN plus FastRelax: combine sequence sampling and structural relaxation; computational scores need independent evaluation.
Research platform · Research platform and methodsMerged
Research workflow budget history
Evidence date:
Preserve prior actions and failures when research budgets change.
Constraints & next steps
- Goal
- Preserve prior actions and failures when research budgets change.
- Constraints
- Budget-history mechanisms are merged; research goals are not established as completed.
- Next step / knowledge feedback
- Validate history continuity and recovery boundaries in applicable workflows.
Tool · Research workflow records
Research platform · Methods and system validationMerged
Capability contracts and validation mechanisms
Evidence date:
Connect capability contracts, execution records and engineering checks so coverage gaps can be identified.
Constraints & next steps
- Goal
- Connect capability contracts, execution records and engineering checks so coverage gaps can be identified.
- Constraints
- Validation mechanisms are merged; self-evolution and scientific effectiveness are not established.
- Next step / knowledge feedback
- Check contract coverage and validation consistency in real tasks.
Tool · Capability validation
Research platform · Research platform and optimization workflowsMerged
Improve code reuse checks
Evidence date:
Avoid constructing unused information for name queries while preserving validation boundaries.
Constraints & next steps
- Goal
- Avoid constructing unused information for name queries while preserving validation boundaries.
- Constraints
- Code merge was checked; historical review records remain retained. Merge does not establish deployment or scientific effectiveness.
- Next step / knowledge feedback
- Track deployment, runtime validation and applicability separately from merge.
Research platform · Research platform and optimization workflowsMerged
Repair validation entry resolution
Evidence date:
Resolve shared validation tools correctly for an explicitly selected source scope.
Constraints & next steps
- Goal
- Resolve shared validation tools correctly for an explicitly selected source scope.
- Constraints
- Code is merged; this does not establish delivery of every downstream task.
- Next step / knowledge feedback
- Complete acceptance before moving to the next stage.
Personal engineering · Personal operations and reliabilityMerged
Calendar operations and readback
Evidence date:
Improve bounded queries, duplicate checks and operation readback while keeping incomplete results unknown.
Constraints & next steps
- Goal
- Improve bounded queries, duplicate checks and operation readback while keeping incomplete results unknown.
- Constraints
- Code is merged; device deployment remains unconfirmed.
- Next step / knowledge feedback
- Complete acceptance before moving to the next stage.
Personal engineering · Personal operations and reliabilityMerged
Agent cancellation and receipts
Evidence date:
Correct cancellation, approval ordering and receipt validation; preserve uncertain operations without replay.
Constraints & next steps
- Goal
- Correct cancellation, approval ordering and receipt validation; preserve uncertain operations without replay.
- Constraints
- Code is merged; runtime validation remains pending.
- Next step / knowledge feedback
- Complete acceptance before moving to the next stage.
Personal engineering · Personal operations and reliabilityBlocked
Recovery after interrupted installation
Evidence date:
Recovery from an interrupted first installation needs explicit ownership rules and an operator decision.
Constraints & next steps
- Goal
- Recovery from an interrupted first installation needs explicit ownership rules and an operator decision.
- Constraints
- Decision pending; this does not imply a failure in running services.
- Next step / knowledge feedback
- Complete acceptance before moving to the next stage.
Research platform · Research platform and optimization workflowsBlocked
Property prediction method research
Evidence date:
Develop research changes around contextual use of property prediction methods and supporting evidence.
Constraints & next steps
- Goal
- Develop research changes around contextual use of property prediction methods and supporting evidence.
- Constraints
- Research is blocked; scientific validation remains incomplete.
- Next step / knowledge feedback
- Resolve the blockers and complete acceptance.
Algorithm · In-context learning: predict conditioned on reference examples; applicability requires explicit controlled evaluation.
Research platform · Research platform and optimization workflowsMerged
Test collection and replay qualification
Evidence date:
Improve test collection and qualification of retained replay records so evidence scope remains traceable.
Constraints & next steps
- Goal
- Improve test collection and qualification of retained replay records so evidence scope remains traceable.
- Constraints
- Code merge was checked; historical review records remain retained. Merge does not establish deployment or scientific effectiveness.
- Next step / knowledge feedback
- Track deployment, runtime validation and applicability separately from merge.
Research platform · Research platform engineeringMerged
Keep engineering checks relevant
Evidence date:
Small changes could wait for unrelated tests while independent work blocked each other. Select checks by the changed scope and run independent checks separately.
Constraints & next steps
- Goal
- Small changes could wait for unrelated tests while independent work blocked each other. Select checks by the changed scope and run independent checks separately.
- Constraints
- Actual queue time and overall speedup remain unmeasured; no timing target is claimed.
- Next step / knowledge feedback
- Actual queue time and overall speedup remain unmeasured; no timing target is claimed.
Research platform · Research platform engineeringMerged
Reduce false expiry during waits
Evidence date:
A service still doing work could be treated as expired while shared state was being updated. Record state at a consistent point and keep liveness signals independent of slower shared writes.
Constraints & next steps
- Goal
- A service still doing work could be treated as expired while shared state was being updated. Record state at a consistent point and keep liveness signals independent of slower shared writes.
- Constraints
- A code repair does not establish sustained production reliability; runtime acceptance is separate.
- Next step / knowledge feedback
- A code repair does not establish sustained production reliability; runtime acceptance is separate.
Research platform · Research platform engineeringMerged
Hand completed results over reliably
Evidence date:
Completed results could be rejected at handoff because their format was interpreted differently. Use the actual result format and check provenance and completeness before accepting it.
Constraints & next steps
- Goal
- Completed results could be rejected at handoff because their format was interpreted differently. Use the actual result format and check provenance and completeness before accepting it.
- Constraints
- Isolated replay is not production acceptance and establishes neither model accuracy nor scientific effectiveness.
- Next step / knowledge feedback
- Isolated replay is not production acceptance and establishes neither model accuracy nor scientific effectiveness.
Research platform · Research tool reliabilityMerged
A real launch before readiness
Evidence date:
Avoid treating a browser environment as ready when it cannot open a page.
Constraints & next steps
- Goal
- Avoid treating a browser environment as ready when it cannot open a page.
- Constraints
- The repair is merged; passing checks does not establish deployment across all runtime environments.
- Next step / knowledge feedback
- Verify launch and page-read failure boundaries in actual runtime environments.
Tool · Browser automation
Research platform · Research workflow traceabilityMerged
A checklist with evidence
Evidence date:
Make task state, evidence, blockers and next actions inspectable together.
Constraints & next steps
- Goal
- Make task state, evidence, blockers and next actions inspectable together.
- Constraints
- The checklist is a view of underlying records; it cannot rewrite state or establish that every task is complete.
- Next step / knowledge feedback
- Keep checking that missing evidence, blockers and next actions agree with the underlying records.
Skill · Evidence checking
Personal engineering · Personal tool engineeringMerged
Diagnostics from field shape
Evidence date:
Distinguish a non-object document from missing required fields so schema errors are easier to locate.
Constraints & next steps
- Goal
- Distinguish a non-object document from missing required fields so schema errors are easier to locate.
- Constraints
- Report only fixed-field presence and JSON types, without field values; a merge does not establish deployment or operational health.
- Next step / knowledge feedback
- Independently check malformed inputs and diagnostic output boundaries; deployment and real-environment behavior still need separate verification.
Research platform · Ongoing developmentMerged
Clarifying research tool responsibilities
Evidence date:
Make design, validation and delivery tools easier to find while keeping their responsibilities clear.
Constraints & next steps
- Goal
- Make design, validation and delivery tools easier to find while keeping their responsibilities clear.
- Constraints
- Merged refers only to source integration of tool responsibilities; scientific applicability and live service acceptance require their own checks.
- Next step / knowledge feedback
- Check actual use against each tool’s qualification conditions and retain earlier failures and limits.
Research platform · Ongoing developmentDeveloping
Making overnight work more reliable
Evidence date:
Preserve evidence through storage or database interruptions and make scheduled monitoring recover or fail visibly.
Constraints & next steps
- Goal
- Preserve evidence through storage or database interruptions and make scheduled monitoring recover or fail visibly.
- Constraints
- Recovery and backups require real readback; the proposed change remains unintegrated.
- Next step / knowledge feedback
- Reconcile independent checks and deployment conditions while retaining failure diagnostics.
Research platform · Ongoing developmentPlanned
Defining data collection before use
Evidence date:
Prepare a traceable data collection plan for later property-prediction research.
Constraints & next steps
- Goal
- Prepare a traceable data collection plan for later property-prediction research.
- Constraints
- Only a pending registration and collection plan exist; no usable dataset, fitted model or qualified benchmark is delivered.
- Next step / knowledge feedback
- Keep this as an open proposal pending qualified source data and further confirmation.
Research platform · Ongoing developmentMerged
Preserving validation history during migration
Evidence date:
Preserve original check obligations and accepted data through tool migrations and interrupted retries.
Constraints & next steps
- Goal
- Preserve original check obligations and accepted data through tool migrations and interrupted retries.
- Constraints
- Scope is the engineering implementation for retaining check history and interrupted publication; it does not establish deployment of all services or scientific qualification.
- Next step / knowledge feedback
- Continue checking provenance, resume positions and final content in later migrations.
Research platform · Ongoing developmentDeveloping
Making request recovery traceable
Evidence date:
Let clients recover original request identity and execution evidence before retrying or resuming.
Constraints & next steps
- Goal
- Let clients recover original request identity and execution evidence before retrying or resuming.
- Constraints
- Retain an unknown state when evidence is incomplete and avoid automatic duplicate submissions.
- Next step / knowledge feedback
- Resolve missing validation while keeping recovery distinct from a new execution.
Personal engineering · Evidence-aware engineeringPublished
Keep two calendar boundaries explicit
Evidence date:
Distinguish a completed Beijing workday from a genuinely partial UTC activity snapshot.
Constraints & next steps
- Goal
- Distinguish a completed Beijing workday from a genuinely partial UTC activity snapshot.
- Constraints
- Publication covers the text and calendar-boundary display; predictions remain pending and existing model authorization and native-result acceptance requirements remain in force.
- Next step / knowledge feedback
- Retain prediction-pending labels and add results only when their original evidence can be checked.
Research platform · Evidence-aware research toolsMerged
Separate comparison evidence from presentation
Evidence date:
Keep comparison context explicit while showing compact views and preserving access to full records.
Constraints & next steps
- Goal
- Keep comparison context explicit while showing compact views and preserving access to full records.
- Constraints
- Source integration is verified. Recorded validation is not independently replayed here and proves neither molecular activity nor every proposal-bearing flow.
- Next step / knowledge feedback
- Retain provenance and limitations; evaluate new cases within their own declared context.
Papers and methods. Research ↗ Capability catalog ↗