Project cases and verification
Business tasks, my responsibilities, implementation choices, and verification results for each project.
Annual-report sentiment analysis delivery
What stands out:Extract topic-specific sentences from annual reports and export sentence-level results in a Windows workflow the client could run.
Paid client delivery | Client materials confidential- My contribution
- I packaged sentence-level sentiment analysis in a Windows GUI with annual-report / ZIP batch input and Excel / CSV export. The single-company version came first, followed by multi-company batching. The client processed and exported results using the guide, then accepted and paid for the delivery.
- Product scope
- The tool supports investment-research information organization and analysis. Final investment decisions remain with the user.
- Materials management
- Client identities, original reports, source code, communications, acceptance records, and payment materials are handled confidentially.
- Delivery outcome
- The client operated the tool and exported structured results before acceptance and payment.
Collaboration Radar
What stands out:Separate proposals from confirmed decisions, trace tasks to their sources, and leave unstated deadlines open.
Expert ranking runner-up | Popularity ranking winner- Task context
- Turn specified group chats and meeting notes into a collaboration status view, helping coordinators verify arrangements, dependencies, and unresolved questions.
- My responsibilities
- Led product and technical design: the seven-module status card, model / code responsibilities, and review / export workflows. Used AI-assisted coding.
- Deliverables
- Structured status cards, rule-derived SVG radars, CSV files, and platform field mappings. People verify meaning and commitments; mappings do not establish that data was written to Lark or Notion.
2026 iFLYTEK AI Developer Competition, Intelligent Office Collaboration Assistant Skill Challenge. Track: 1,851 teams; account: jamji. In-person awards on October 24, 2026. View the competition page ↗
Read the full case: input → status cards → source review → export ↗
Expense precheck: evidence completeness
What stands out:Distinguish missing evidence from a policy violation, and explain what to supply before rule review.
Local POC | Evidence checks and human review- Verified
- Original ZIP integrity check passed;
evidence-gatemissing or conflicting evidence returnsFLAGand a human action. All 3 precheck examples and 12 minimal regression tests passed in self-tests. - Evaluation
- Evaluation on 30 samples from the original labeled dataset: decision accuracy 63.33%, precision 87.50%, and recall 82.35% (all self-tested). Decision accuracy was below the self-set 80% threshold.
- API read records
- The ZIP retains
/v1/travel-standardsread results for 12 travel-policy master-data records. - Workflow scope
- Evidence checks before rule-based review,
FLAGrouting, and human collection of missing evidence.
WeRead AI Brain
What stands out:Compare ideas across books around one question, check passages, and remove unsupported links before export.
Open source on GitHub | Bring your own API key- Public sources
- GitHub:yamyeed/weread-ai-brain。
- What it does
- Combines WeRead data, a visual dashboard, AI analysis, and Markdown knowledge export. Skill documentation includes dependencies, commands, and environment variables.
- Prerequisites
- Users provide an API key through local environment variables. The public repository contains no personal keys.
- Usage
- Available as an open-source repository and Skill configuration. The public cross-book walkthrough uses manually written interpretations for source review, correction, and export, without connecting to a personal WeRead account.
Agent Task Evaluation
Check task completion and trace failures through tool calls, file handling, and saved deliverables.
External task records | 5 model aliases · 3 task types · 2 execution environments- User task
- Compound tasks require tools, file processing, and saved results. Evaluation checks the actual artifacts from each step against the task requirements.
- My contribution
- Recorded tool calls, code execution, data analysis, and saved deliverables across five model aliases, three compound task types, and two execution environments.
- Verification
- Compared task requirements, execution records, and screenshots of evaluation artifacts. Located failures in task understanding, tool inputs, execution, saving, or supporting evidence.
- Findings
- Documented context loss, interrupted tool chains, inconsistent artifact formats, and conclusions without supporting data, informing checks, human confirmation, and error recovery.
- Material scope
- This page summarizes methods and failure patterns. Original task materials are confidential, and a complete public redacted package remains unavailable. The records support observations within the specific tasks.
Text2SQL / Data Agent
What stands out:Complete composite join keys along trusted paths, and stop when the path is ambiguous or unauthorized.
Local Agent prototype | Separate schema evaluation- User task
- Turn natural-language data requests into a flow with clarification and validation, while constraining generated SQL before execution.
- My contribution
- Built on a shared training-program LangGraph workflow, maintained the clarification, schema retrieval, SQL generation, and execution flow, and tightened SQL validation and read-only safeguards.
- Verification results
- The September 10, 2026 local safety checks passed 11 offline regressions and two MySQL permission tests. Natural-language SQL generation accuracy and end-to-end task completion remain unverified.
- Data environment
- The Agent prototype uses a local MySQL sample database. The separate schema evaluator uses synthetic SQLite data.
- Open-source contribution
- Contributed Chinese and English explanations of schema selection and join paths to awesome-ai-roadmap. After maintainer review and revisions, PR #17 was merged on September 29, 2026.
- Follow-up evaluation
- Built a Python / SQLite evaluator with 8 synthetic tables, 9 test inputs, and 34 automated tests for join-key completion, permission constraints, path ambiguity, and insufficient information.
- Verification method
- Compare column recall before / after completion, complete query policies, and fixed reference SQL results. Require clarification for ambiguous joins and retain refusal branches when no path or permission is available.
Scientific Nutrition Advisor
What stands out:Suggest takeaway options around budget and preferences, clarify missing details, and flag situations that need professional guidance.
Publicly released | Public platform pages- Public sources
- Chinese edition 、 English edition 、 iFLYTEK award announcement 。
- Download statistics
- Cumulative downloads for each language edition are available on its public ClawHub page, using the platform's distribution metric.
- Health guidance
- Product guidance includes referral paths for special populations, medical conditions, and medical advice. The tool supports health management.
AI Reading Coach
What stands out:Change the work goal to see a different reading priority for the same technical book.
AI reading product | Open source | 40 local-preview tests passed- Product task
- Turn one-off AI questions into a stateful, trackable reading workflow: tailor a book map to the reader's background, difficulties, and goals; guide each chapter; then check understanding through application questions.
- My contribution
- Specified data models, chapter parsing, reading batches, progress states, generation constraints, fallback behavior, and acceptance criteria in structured requirements readable by AI tools, then developed the UI prototype.
- Product boundaries
- The HTTP Agent interface supports SSE, NDJSON, JSON, and plain text. The public demo uses a local streaming engine. The developer panel exposes the generation source, estimated tokens, and timing.
- Validation
- The November 3, 2026 local preview passed 40 tests, type checking, and the build. Tests found and helped fix a bug that ignored supplied chapter state when creating a book.
- Current status
- The open-source application includes a bookshelf, chapter progress, and an Agent integration interface. This offline demo matches goal terms to chapter titles. Generated content and questions demonstrate the flow; target-user outcomes remain unverified.
PersonalOS Tool-Calling Agent
Local application module | JSON Schema | SQLite session history- User task
- Let users ask about knowledge and articles, with an assistant invoking retrieval tools and retaining reviewable conversations.
- Choice & cost
- I bounded visible tools, history, and loop rounds to control execution. Keyword-based selection may omit a tool and trimmed history may lose context, creating a tradeoff between task coverage and execution budget.
- Validation & next step
- Existing records support inspection of tool contracts, conversation storage, and execution limits. Next: test multi-turn follow-ups and retrieval failures for missing information and recovery; runtime success alone does not establish answer quality.
Expand implementation & evidence
- User task
- In a local-first, multi-module workflow system, the model selects knowledge search, full-text search, or article lookup tools and makes bounded multi-round calls.
- My contribution
- Tools are defined with JSON Schema. The model proposes a call; the system validates its arguments, executes an approved tool, and returns the result to the conversation.
- Context boundaries
- Filter visible tool schemas by question keywords. Expose only knowledge and full-text search by default, and limit tool-result length and conversation history.
- Execution boundaries
- Keep at most the last 10 messages and cap the Agent loop at 10 rounds. Store conversations and used-tool records in local SQLite.
- Reviewable interfaces
- Authenticated APIs list tools and create, read, or delete conversations, supporting session recovery and tool-call review.
- Module checks
- Twenty targeted local regression checks passed on October 29, 2026, covering tool loops, failure handling, confirmation and cancellation, context limits, and retrieval sources.
- Public walkthrough scope
- Isolated checks use three public documents and cover empty retrieval, cancellation, expiry, and repeat confirmation. Drafts combine retrieved excerpts. The complete main-application model and knowledge workflow still needs end-to-end verification.
Spring Boot Agent Gateway
What stands out:Use tool traces to diagnose invalid arguments, blocked tool calls, and timeouts in a Java integration.
Personal engineering project | JDK 21 + Maven | 6/6 tests passed- My contribution
- Built with Java 21 / Spring Boot 3: an Agent Chat API, bounded loop, allowlisted Tool Registry, JSON Schema validation, tool traces, session storage, and minimal RAG.
- Tool management
- Registers
knowledge_searchand structuredread_only_query; execution is controlled through an allowlist, JSON Schema, timeouts, and bounded retries. - Validation
- Compiled on JDK 21.0.12 and Maven 3.9.16. Six tests across four classes passed, covering tool execution, RAG chunking/retrieval, and the application Agent loop.
- How to run
mvn test; local checks use H2 and an in-memory adapter. MySQL, Redis, and online-model integrations require further verification.- Extension points
- The default HashEmbeddingPort validates the service flow. The EmbeddingPort interface supports replacement and retrieval evaluation on specific datasets.
Inspect MCP protocol exchanges, source and reproduction steps ↗
Media Extract
What stands out:Use existing text layers or subtitles first, keep source references, and retry only unfinished items.
Multimedia processing | Source traceability | Resumable batchesRoutes materials through text layers, embedded subtitles, OCR, or speech recognition, producing readable text, structured metadata, and source artifacts.