Skip to content

Pelis Agent Factory Advisor #108

Pelis Agent Factory Advisor

Pelis Agent Factory Advisor #108

---
description: Daily advisor that analyzes the repository for opportunities to add, enhance, or improve agentic workflows based on Pelis Agent Factory patterns
on:
schedule: daily
workflow_dispatch:
permissions:
contents: read
actions: read
issues: read
pull-requests: read
discussions: read
tools:
agentic-workflows:
bash:
- "cat"
- "find"
- "ls"
- "grep"
cache-memory: true
github:
toolsets: [context]
sandbox:
agent:
version: v0.25.29
network:
allowed:
- "github.github.io"
safe-outputs:
threat-detection:
enabled: false
create-discussion:
title-prefix: "[Pelis Agent Factory Advisor] "
category: "general"
timeout-minutes: 30
steps:
- name: Fetch Pelis Agent Factory Docs
id: fetch-docs
run: |
set -o pipefail
BASE="https://github.github.io/gh-aw"
OUTFILE="${GITHUB_WORKSPACE}/.pelis-agent-factory-docs.txt"
: > "$OUTFILE"
for PATH_SUFFIX in \
"/blog/2026-01-12-welcome-to-pelis-agent-factory/" \
"/introduction/overview/" \
"/guides/workflow-patterns/" \
"/guides/best-practices/"; do
echo "### ${BASE}${PATH_SUFFIX}" >> "$OUTFILE"
curl -sf "${BASE}${PATH_SUFFIX}" \
| python3 -c "import sys,html,re;t=sys.stdin.read();print(html.unescape(re.sub('<[^>]+>','',t))[:3500])" \
>> "$OUTFILE" 2>/dev/null \
|| echo "(not found)" >> "$OUTFILE"
echo "" >> "$OUTFILE"
done
- name: Fetch Agentics Patterns
id: fetch-agentics
run: |
set -o pipefail
curl -sf "https://raw.githubusercontent.com/githubnext/agentics/main/README.md" \
| head -c 4000 > "${GITHUB_WORKSPACE}/.agentics-patterns.txt" \
|| echo "(not available)" > "${GITHUB_WORKSPACE}/.agentics-patterns.txt"
- name: Compute Content Hashes
id: content-hashes
run: |
{
sha256sum "${GITHUB_WORKSPACE}/.pelis-agent-factory-docs.txt"
sha256sum "${GITHUB_WORKSPACE}/.agentics-patterns.txt"
} | sha256sum | cut -d' ' -f1 > "${GITHUB_WORKSPACE}/.content-hash.txt"
- name: Summarize Existing Workflows
id: workflow-summaries
run: |
{
echo "# Workflow Summaries (name | description | triggers)"
find .github/workflows -name "*.md" -type f | sort | while IFS= read -r f; do
name=$(basename "$f" .md)
desc=$(grep -m1 "^description:" "$f" 2>/dev/null | sed 's/^description: *//' | cut -c1-100)
triggers=$(grep -E "^ (schedule|workflow_dispatch|pull_request|push|issues|workflow_run|issue_comment|release|slash_command):" "$f" 2>/dev/null \
| sed 's/^ //' | tr -d ':' | tr '\n' ',' | sed 's/,$//')
printf "%-45s | %-100s | %s\n" "$name" "${desc:-(no description)}" "${triggers:-(none)}"
done
} > "${GITHUB_WORKSPACE}/.workflow-summaries.txt"
- name: Collect Repo Structure
id: repo-structure
run: |
{
echo "=== Root files ==="
ls -la
echo ""
echo "=== Tests ==="
ls -la tests/ 2>/dev/null || echo "(no tests/)"
echo ""
echo "=== Scripts ==="
ls -la scripts/ 2>/dev/null || echo "(no scripts/)"
} > "${GITHUB_WORKSPACE}/.repo-structure.txt"
---
# Pelis Agent Factory Advisor
You are an expert advisor on agentic workflows specializing in Pelis Agent Factory patterns. Your mission: identify the top opportunities to add, enhance, or improve agentic workflows in this repository.
> **Batch all independent operations into a single turn.** Never read files one at a time when you can read them in parallel.
## Phase 1: Learn Patterns (cache-gated)
**First turn**: read `.content-hash.txt` and `.workflow-summaries.txt` in parallel (both are always read on every run).
Compare `.content-hash.txt` to cached `pelis_docs_hash` in cache-memory.
- **Cache hit** (hash unchanged): skip reading doc files; use cached patterns and proceed directly to Phase 2.
- **Cache miss**: read `.pelis-agent-factory-docs.txt` and `.agentics-patterns.txt` in a single parallel batch, then update `pelis_docs_hash` and store a bullet-point summary of key patterns in cache-memory (`pelis_patterns_summary`).
Key things to extract from docs: workflow templates, safe-outputs patterns, caching strategies, permission models, integration patterns.
## Phase 2: Analyze Repository
All repository context is pre-computed — avoid reading individual workflow files unless you need specific configuration details.
**In one turn**, do all of the following in parallel:
1. `agentic-workflows status` — check recent run health
2. `agentic-workflows audit` — check security/config issues
3. `bash:cat .repo-structure.txt` — root files, tests, scripts
The pre-computed `.workflow-summaries.txt` (read in Phase 1 first turn) gives you a one-line inventory of every workflow (name | description | triggers). Use it as your primary inventory — **do not read individual workflow `.md` files** unless a specific recommendation requires detailed configuration review.
Assess automation coverage: what triggers exist, what's missing, where there are gaps.
## Phase 3: Identify Opportunities
Identify top opportunities across these categories (focus on highest value for a security/firewall tool):
- **Missing workflows**: security automation, test quality, release automation, documentation, monitoring
- **Enhancement opportunities**: caching, triggers, tool utilization, error handling
- **Integration opportunities**: chaining workflows, shared state, event-driven patterns
For each opportunity: Impact (H/M/L) · Effort (H/M/L) · Risk (H/M/L).
Priority: P0=High impact+Low effort, P1=High impact+Medium effort, P2=Medium, P3=Nice-to-have.
## Phase 4: Report
Create a discussion using `create_discussion` with:
1. **📊 Executive Summary** — 2–3 sentences on maturity and top opportunities
2. **📋 Workflow Inventory** — Table from `.workflow-summaries.txt`: `| Workflow | Purpose | Trigger | Assessment |`
3. **🚀 Recommendations** — Grouped P0→P3, each with: What / Why / How / Effort
4. **📈 Maturity Assessment** — Current/Target level (1–5), gap analysis
5. **📝 Cache Update** — Update cache-memory with observed patterns and items to track next run
## Guidelines
- Specific and actionable — every recommendation must be implementable
- Security-focused — this is a firewall tool; prioritize security-relevant automations
- Ruthlessly prioritize — top 5 wins over exhaustive lists
- Use cached knowledge — avoid re-reading docs and files already summarized