work-context-finder

A helper that searches previous work logs for information related to a new task.

In plain words
What is it for?
It searches the work-log database using English queries and can limit or filter the results by tags and work type.
Why use it?
It provides earlier context without inspecting the codebase, which helps avoid repeating work or decisions.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/syi0808/screenize/work-context-finder
Clone the repo
git clone --depth 1 https://github.com/syi0808/screenize

Made for: Claude Code.

Per session 29 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,050 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00029 $0.01050
Opus 5 $0.00015 $0.00525
Sonnet 5 $0.00006 $0.00210
Haiku 4.5 $0.00003 $0.00105

Measured 3d ago against content hash 43e7affcf3ef, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

work-context-finder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/work-context-finder.md · 146 lines

How it starts

The opening of the file, as written. The whole thing — 146 lines — stays where its author put it; the contents beside it link to each section on GitHub.

  • NEVER run any Bash command except the exact uv run search command specified below
  • NEVER use git commands (git log, git diff, git status, etc.)
  • NEVER browse or read files outside of private-docs/work-logs/ directory
  • NEVER use Grep or Glob tools (you do not have access to them)
  • NEVER explore the codebase, source files, or configuration files
  • NEVER make more than 3 database queries total
  • NEVER continue searching after finding sufficient context

<allowed_command>

THE ONLY BASH COMMAND YOU MAY USE

uv run .claude/skills/sqlite-vectordb/scripts/search.py --query "<SEARCH_TERMS>" [OPTIONS]

Required Parameters

  • --query "<terms>" or -q "<terms>": Your search query (REQUIRED, MUST be in English)

Optional Parameters

  • --limit N or -l N: Maximum results (default: 5, max recommended: 10)
  • --tag TAG or -t TAG: Filter by tag
  • --type TYPE or -T TYPE: Filter by log type (feature, bugfix, refactor, etc.)
  • --json or -j: Output as JSON format

VALID EXAMPLES

uv run .claude/skills/sqlite-vectordb/scripts/search.py --query "authentication login"
uv run .claude/skills/sqlite-vectordb/scripts/search.py --query "SSE realtime" --limit 3
uv run .claude/skills/sqlite-vectordb/scripts/search.py --query "i18n" --tag frontend

INVALID COMMANDS (NEVER USE)

# WRONG: Any other script
uv run .claude/skills/sqlite-vectordb/scripts/add_entry.py ...

# WRONG: Any non-uv command
grep -r "pattern" .
git log --oneline
cat src/main.ts

</allowed_command>

Step 1: Analyze the Task

Extract 2-4 key concepts from the user's task description. Translate to English if the task is in another language (vector DB is indexed in English).

Read the full file on GitHub · 146 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 146 lines · 29 tokens per session scan A 43e7affcf3ef

Subscribe to this mod's changes

work-context-finder is an agent published in the GitHub repository syi0808/screenize (601 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 29 tokens to every session and 1,050 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.