Pre-launch — Waitlist open

Talent management
for AI agents.

SkillSpec helps teams building AI agents assess, track, and improve the skills their agents use in production, so capability doesn't drift and good agents don't stay good by accident.

SKILLSPEC — WORKSPACE DASHBOARD
Skills 24
Avg Score 78
Directives 3
Document Analysis
92 ↑ 4
Code Generation
81 ↑ 7
Email Drafting
74
Data Extraction
63 ↓ 5
Meeting Summary
58 ↓ 3
Image Description
34 ↓ 12
PENDING DIRECTIVES
⚡ IMPROVEMENT
Image Description scoring low on structured output. Proposed SKILL.md update targets format consistency.
30-DAY TREND

You deployed your agents.
Now you have to manage them.

Most teams can deploy agents, but they can't systematically manage capability over time. They don't know which skills are strong, which are weak, or whether a change to a SKILL.md actually helped. As the number of agents grows, capability management turns into guesswork.

No assessment

You can review outputs, but you don't have a reliable view of which skills are performing well, underperforming, or drifting over time.

No development loop

When a skill underperforms, improvement is manual: read feedback, guess at the cause, edit SKILL.md, redeploy, and hope the change helps.

No performance history

There's no clear record of whether a skill improved, regressed, or stayed stagnant. Without that history, every change is a guess.

Some of the biggest capability problems are easy to miss until they've already affected results.

Silent failure

Skills that never activate

A skill can be perfectly written but never trigger if its description doesn't match how users phrase their requests. Your carefully crafted SKILL.md sits idle — and you'd never know.

# Never triggers because...
description: "Generates quarterly
  financial reconciliation"
# User asks:
"Can you check last quarter's
  numbers match up?"
Hidden drift

Skills that fall out of date

A skill that used to work can become less effective as prompts, tooling, or team expectations change. Without ongoing assessment, that drift stays invisible.

# Last updated for an older workflow...
tool: "legacy-reporting"
# But base.md contains:
Output omits the new review checklist.
Team now expects structured evidence.

A development loop for
agent capability.

SkillSpec gives teams a structured way to assess skills, improve them, and verify what actually got better.

01

Assess

Capture structured feedback on how a skill performed in real work, not just isolated tests.

02

Diagnose

Identify recurring patterns in where a skill is helping, failing, or degrading across repeated use.

03

Improve

Generate targeted SKILL.md changes, review them, and apply the improvements worth keeping.

04

Verify

Track whether the change actually improved outcomes over time, so development is evidence-based instead of anecdotal.

Everything you need to manage
agent capability at scale.

See which skills are strong, weak, and drifting

Track capability at the skill level across agents, with trends, targets, and signals that show where attention is needed.

  • Per-agent, per-skill assessment
  • Trend charts and history over time
  • Gap analysis against targets
  • Signals for stale or underused skills
SKILL PROFICIENCY
Code Generation
88
Email Drafting
74
Data Extraction
61
Image Description
34

Turn feedback into concrete skill improvements

SkillSpec converts recurring feedback into proposed SKILL.md changes, then tracks whether those changes improved results.

  • Pattern analysis across repeated feedback
  • Targeted suggestions with clear rationale
  • Unified diff view of exact SKILL.md changes
  • Before-and-after effectiveness tracking
DIRECTIVE — IMAGE DESCRIPTION
⚡ HIGH IMPACT
Add structured output format requirements to prevent inconsistent descriptions.
  ## Output Requirements
- Describe images clearly and concisely.
+ Always structure output as:
+ 1. Subject identification
+ 2. Key visual attributes
+ 3. Contextual description
+ 4. Accessibility alt-text (max 125 chars)
 
  ## Tone

Let agents drive their own improvement

SkillSpec gives agents structured visibility into their own skills, weak spots, and approved improvement directives, so they can adapt and improve within the guardrails you control.

  • Agent-readable skill context
  • Approved directives available in execution
  • Workspace-controlled automation and self-review
  • Faster improvement without a human bottleneck
AGENT API
GET /api/v1/agent/context
{
  "skill": "image-description",
  "proficiency": 34,
  "trend": "declining",
  "pending_directives": 1,
  "auto_apply": true
}

Create a shared standard for agent capability

Manage skills as a team with shared libraries, visibility controls, and a common view of what good looks like.

  • Workspace model with roles
  • Private and shared skill visibility
  • Shared skill libraries and categories
  • Portfolio-level oversight for admins
WORKSPACE — TEAM
SH
Sarah Henderson
sarah@company.com
ADMIN
MK
Marcus Kim
marcus@company.com
MANAGER
JP
Jasmine Patel
jasmine@company.com
MEMBER
TC
Tom Chen
tom@company.com
MEMBER

Build a better system for
managing agent skills.

Join the waitlist to get early access and help shape how teams assess and improve agent capability in production.