Tech Radar Context Agent Skill

tech-radar-context

Overview

Surface Tikal's opinion on the technologies relevant to the current prompt so

a tech stack choice is informed by the Israeli Tech Radar. This skill works like

team-discover, but its search surface is the Tikal Tech Radar dataset

(radar.json) instead of the team CDR index: it extracts candidate technologies

from the prompt, matches them against radar blips, and injects a compact

Tech Radar Context table (ring, quadrant, Tikal's "Why?" opinion) plus

Tikal-aligned alternatives for anything on Stop.

The radar has four quadrants — DevOps, Backend, AI/ML, Web/Mobile —

and four adoption rings:

| Ring | Meaning | Guidance |

|------|---------|----------|

| Try | New stuff that on the surface seems good (good press, new solution) | Explore / evaluate; not yet endorsed for production use |

| Start | A good solution more companies should use; if in beta, active progress and contribution | Recommend adopting on new projects |

| Keep | Stable release (non-beta) with major supporter acceptance (large community, used by corporates) | Recommend by default for current & new work |

| Stop | Items we recommend companies stop using — better alternatives exist | Warn against; recommend a Keep/Start alternative |

Each blip's description embeds an HTML <p>Why?</p> block followed by a

<p>Description</p> block. The Why? text carries Tikal's explicit stance and

rationale — that is the opinion to surface. A technology may appear more than

once (different quadrants) with different rings; report each relevant placement.

When to Use

Model-invoke this skill whenever the prompt involves **choosing or evaluating

technology**, for example:

• Selecting a framework, library, database, message broker, or cloud tool.
• Comparing options ("X vs Y", "should we use Z").
• Designing a system, service, or pipeline where stack decisions are implied.
• Reviewing an existing stack for modernization or replacement.

Do not invoke it for pure business/product questions with no technology

selection, or when the user explicitly says to ignore the radar.

Manual invocation:

/tech-radar-context               # inject radar context for the current prompt
/tech-radar-context redis vs kafka

Core Process

Step 1: Extract Candidate Technologies

From the current prompt (or the description provided by an invoking skill),

extract every named or clearly implied technology: languages, frameworks,

libraries, databases, brokers, CI/CD tools, cloud services, AI/LLM tooling,

build tools, etc. Normalize obvious aliases (e.g. "postgres" → "PostgreSQL",

"k8s" → "Kubernetes", "GH Actions" → "GitHub Actions").

If the prompt implies a category without naming a product (e.g. "we need a

vector database", "pick a Python web framework"), treat the category as a query

and surface the radar's recommended options in that space.

Step 2: Load and Query the Radar Dataset

Execute the deterministic search helper script (relative to this skill directory):

POSIX (bash + jq):

bash scripts/radar-search.sh <tech1> [tech2 ...]

Windows (PowerShell):

pwsh scripts/radar-search.ps1 <tech1> [tech2 ...]

Or for JSON output (add --json for bash, -Json for PowerShell).

The script handles alias mapping (k8s → Kubernetes, postgres → PostgreSQL, gh actions → GitHub Actions, etc.), matches against resources/radar.json, extracts Tikal's <p>Why?</p> opinion, and formats the markdown table automatically.

If a technology appears in multiple placements with conflicting rings (e.g.,

Node.js is both DevOps: Stop and DevOps: Keep), the script detects and

flags it with a Conflicting Guidance note.

Schema of resources/radar.json:

{
  "title": "Explore the Tech Radar",
  "quadrants": ["DevOps", "Backend", "AI/ML", "Web/Mobile"],
  "rings": ["Try", "Start", "Keep", "Stop"],
  "blips": [
    {
      "name": "FastAPI",
      "quadrant": "Backend",
      "ring": "Keep",
      "description": "<p>Why?</p>\n<p>...Tikal's opinion...</p>\n<p>Description</p>\n<p>...</p>",
      "isNew": false
    }
  ]
}

Step 3: Match Blips

For each candidate from Step 1, find matching blips by name (case-insensitive,

alias-aware, allowing minor version suffixes like "Airflow 2" / "Airflow 3" and

partial matches like "Redux" → "Redux Toolkit"). A candidate may match multiple

blips across quadrants — keep them all.

For category queries (Step 1), select the strongest radar recommendations in

that space: prefer Keep/Start blips in the matching quadrant, and note any

Stop blips as things to avoid.

Step 4: Extract Tikal's Opinion

For every matched blip, parse the description HTML:

• The text inside the <p>Why?</p> block (up to the next <p>Description</p>)

is Tikal's opinion / rationale — the primary signal.

• The <p>Description</p> block is neutral background — use only if helpful.

Strip HTML tags to plain text and condense the "Why?" to one or two sentences

for the context table (quote it more fully when the ring is Stop or when the

user is directly weighing that technology).

Step 5: Recommend Alternatives for Stop

When a candidate matches a Stop blip (or is a legacy technology the radar

clearly discourages), select Tikal-aligned replacements from the **same

quadrant** on Keep or Start, guided by the "Why?" text. Common examples the

dataset supports:

• Airflow 2 → Airflow 3 / Dagster
• Create React App → Vite / Next.js
• Jenkins → GitHub Actions / GitLab CI / Tekton
• requirements.txt / Poetry → uv
• Moment.js / Luxon → day.js / date-fns
• Redux / Redux Toolkit → Zustand / TanStack Query
• Ant Design / Styled Components → Tailwind CSS / shadcn/ui / Radix UI

Do not hardcode substitutions beyond what the loaded dataset supports — derive

alternatives from the radar's actual Keep/Start blips in that quadrant.

Step 6: Inject Tech Radar Context (Output Contract)

Emit a Tikal Tech Radar Context section in the visible response, before the

task answer, so downstream reasoning is grounded in the radar:

## Tikal Tech Radar Context

| Technology | Quadrant | Ring | Tikal's Opinion (Why?) |
|------------|----------|------|------------------------|
| FastAPI | Backend | Keep | Better alternative to Flask; async, fast, big and growing community. |
| Jenkins | Backend | Stop | Plugin hell + XML config; legacy vs GitHub Actions / GitLab CI / Tekton. |

**Radar guidance**
- ✅ Keep/Start: FastAPI — safe to adopt.
- ⚠️ Stop: Jenkins → consider GitHub Actions, GitLab CI, or Tekton (see Why? above).

_Source: Tikal Israeli Tech Radar (local snapshot) · N technologies matched._

• One row per matched blip (include duplicates across quadrants when relevant).
• Group a short Radar guidance list: safe-to-adopt vs avoid-with-alternatives.
• Add a _Source_ line noting the data came from the local snapshot and

how many technologies matched.

If no candidate technology matches any blip, state that plainly with an empty

table and a _Source_ line (e.g. _Source: … · 0 technologies matched._) —

do not fabricate radar placements.

Failure Handling

• Local resources/radar.json missing/unparseable → emit an empty context

table noting the radar was unavailable, and continue; never block the

user's task.

• No matches → empty table + 0 technologies matched line.

Red Flags

• Fabricating a ring, quadrant, or "Why?" opinion for a technology that is not

in the loaded dataset — report 0 matches instead.

• Blocking on the live fetch instead of falling through to the local snapshot.
• Hardcoding Stop→alternative substitutions not backed by the loaded radar.
• Reporting the neutral <p>Description</p> text as Tikal's opinion — the

opinion lives in the <p>Why?</p> block.

• Ignoring duplicate blips: the same technology can sit in different quadrants

with different rings; surface each relevant placement.

• Treating the skill load as the work — the Core Process must actually run and

produce the Tech Radar Context table.

Verification

• Candidate technologies were extracted from the prompt (or category queries

formed when no product was named).

• The radar dataset was loaded from resources/radar.json.
• A Tikal Tech Radar Context table was produced with columns Technology /

Quadrant / Ring / Tikal's Opinion (Why?), with the opinion sourced from the

<p>Why?</p> block.

• Stop-ring matches include Tikal-aligned Keep/Start alternatives from the

same quadrant, derived from the dataset.

• A _Source_ line reports live-vs-snapshot and the match count; a no-match run

yields an empty table plus 0 technologies matched rather than fabricated data.

Configuration

• Local source: resources/radar.json (bundled full radar snapshot; the only source the search script uses).
• Canonical source: https://www.tikalk.com/radar/ (for periodic manual snapshot regeneration; not fetched by the script at runtime).
• Regenerate the snapshot periodically from the canonical Tikal radar dataset to

stay current.