Sharpen discoverability against same-name agent observability projects

[luoyuctl]

Background

GitHub search for agenttrace shows multiple adjacent projects using the same or very similar name. The current repository is positioned as a local TUI and report generator for AI coding agent session history, which is narrower than general agent SDK observability or web dashboards.

Evidence

• GitHub repository search for agenttrace returns several same-name or same-case projects, including tensorstax/agenttrace, Rxflex/agenttrace, and other AI agent observability/debugging repositories.
• luoyuctl/agenttrace currently has the description: TUI observability for AI coding agents; trace cost, tokens, tool failures, latency, anomalies, health, diffs, and CI gates across local coding-agent sources.
• Repository topics already include strong terms such as ai-agent-observability, ai-coding, token-usage, ai-cost-tracking, claude-code, codex-cli, and gemini-cli.

User value

Users searching for coding-agent session analysis should be able to quickly tell that this project is for local coding-agent logs, TUI triage, and CI/report evidence, not a general tracing SDK or hosted dashboard.

Adoption rationale

Clear discoverability reduces bounce from search results and helps users choose the right tool for local-first session debugging.

Suggested scope

• Review README title/first paragraph, repository description, site H1/supporting copy, and package metadata for a consistent differentiator.
• Prefer wording around local coding-agent session history, terminal/TUI, cost/tokens/time, tool failures, health, and CI gates.
• Keep existing package/repo name unchanged unless maintainers explicitly decide otherwise.

Non-goals

• Do not attack or compare negatively against same-name projects.
• Do not rename the repository or package in this task.
• Do not add platform target language or artificial popularity claims.

Acceptance criteria

• A first-time reader can distinguish this project from general agent tracing SDKs within the first README viewport.
• Repository/site/package surfaces use consistent local-first coding-agent session observability wording.
• Existing topics and metadata remain compatible with current release surfaces.

Suggested lane

lane/growth

Risk

Medium. Copy changes touch public positioning and should stay precise to avoid widening scope beyond what the product actually validates.

Source

Product ecosystem scan on 2026-05-10 using GitHub repository search results and current repo metadata.

[luoyuctl]

Event: Handoff
Actor: Product
Scope: Issue #173
State change: keep lane/growth + status/ready-for-agent
Evidence: GitHub repository search for agenttrace returns several same-name or similar AI agent tracing projects; this repository is specifically a local TUI/report generator for coding-agent session history.
Next owner: Growth

Handoff to: Growth
Reason: The first-viewport positioning should make the local coding-agent session observability use case obvious without expanding product scope.
Expected output: propose copy/metadata updates for README, site, and package surfaces that sharpen the distinction from general tracing SDKs or hosted dashboards.
Acceptance check: first-time readers can distinguish this project as local-first coding-agent session history observability within the first README viewport, and release/install metadata remains accurate.
Context: repo topics already cover ai-agent-observability, ai-coding, token-usage, ai-cost-tracking, claude-code, codex-cli, and gemini-cli.

[luoyuctl]

Event: Evidence / Handoff
Actor: Product
Scope: Issue #173 / PR #156
State change: no label change; keep lane/growth + status/ready-for-agent
Evidence: PR #156 changes README, docs/comparison.md, npm/README.md, and site/llms.txt around the local cost explorer / slow-run debugger positioning. That appears aligned with this issue’s discoverability goal, but PR #156 remains draft + status/needs-human because #174 and #176 are still unresolved.
Next owner: Growth / Human

Decision: Treat PR #156 as the current candidate implementation vehicle for #173, not as a completed fix yet.

Acceptance check: if #156 proceeds, verify the first README viewport and site/package surfaces clearly distinguish agenttrace as local-first coding-agent session history observability, without broadening into a generic tracing SDK or hosted dashboard. Keep #174/#176 blockers separate from this discoverability acceptance.

[luoyuctl]

Event: Decision
Actor: Product
Scope: Issue #173 / PR #156
State change: status/ready-for-agent -> status/needs-human
Evidence: PR #156 is the current candidate implementation vehicle for this discoverability work, touching README, docs/comparison.md, npm/README.md, and site surfaces. It remains draft + status/needs-human because #174 and #176 are unresolved.
Next owner: Human / Growth

Decision: Avoid parallel Growth implementation while #156 remains an active candidate path. Keep #173 in status/needs-human until #156 is either updated/merged, explicitly abandoned, or split so this issue needs a separate allowed-scope PR.

Acceptance check: if #156 proceeds, verify first-viewport README/site/package positioning distinguishes agenttrace as local-first coding-agent session history observability. If #156 is closed without merge or drops this scope, Product can restore status/ready-for-agent with a fresh handoff.

[luoyuctl]

Event: Validation / Decision
Actor: Product
Scope: Issue #173 / PR #156
State change: status/needs-human -> status/ready-for-review
Evidence: Rechecked PR #156 head 3b2b14e. The candidate Growth diff now sharpens first-viewport README positioning around a local cost explorer and slow-run debugger for coding agents; adds docs/comparison.md that distinguishes agenttrace from raw logs, agent UI history, cost-only scripts, CI logs, and hosted dashboards; keeps npm/README.md focused on local-first coding-agent session history without promoting npm as an active install path; and aligns site metadata/hero/source copy with local-first coding-agent session observability. GitHub reports PR #156 mergeStateStatus=CLEAN and CI Test and build SUCCESS.
Next owner: Human / Growth

Decision: #173 product acceptance is satisfied by the current PR #156 scope, pending merge. Keep this issue in ready-for-review rather than ready-for-agent so no parallel Growth PR is created while #156 remains the active implementation path.

Acceptance check after merge: first README viewport, comparison doc, npm README, and site surfaces still distinguish agenttrace as local-first coding-agent session history observability and do not broaden into a generic tracing SDK or hosted dashboard.