This file is indexed.

/usr/share/doc/freeciv-data/README.agents is in freeciv-data 2.5.10-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Overview
========

An agent is a piece of code which is responsible for a certain
area. An agent will be given a specification by the user of the agent
and a set of objects which the agent can control (the production
queue of a city, a city, a unit, a set of units or the whole
empire). The user can be a human player or another part of the code
including another agent. There is no extra interaction between the
user and the agent needed after the agent got its task description.

Examples of agents:
 - an agent which is responsible for moving a certain unit from A to B
 - an agent which is responsible for maximize the food production of a
 city
 - an agent which is responsible for the production queue of a city
 - an agent which is responsible for defending a city
 - an agent which is responsible for a city
 - an agent which is responsible for all cities

An agent may use other agents to accomplish its goal. Such decencies
form a hierarchy of agents. The position in this hierarchy is denoted
by a level. A higher level means more complexity. So an agent of level
n can only make use of agents of level (n-1) or lower. Level 0 defines
actions which are carried out at the server and are atomic actions
(actions which can't be simulated at the client).

By such a definition an agent doesn't have to be implemented in C and
also doesn't have to make use of client/agents/agents.[ch].

The core of an agent consist of two parts: a part which makes
decisions and a part which carries out the decision. The results of
the first part should be made available. An agent lacking the
execution part is called advisor.

An agent should provide a GUI besides the core.

Implementation
==============

The received task description and any decision been made can be saved
in attributes. An agent should not assume anything. This includes
especially: _no magic numbers_. Everything should be settable by the
user.

Use client/agents/agents.[ch] to get informed about certain
events. Don't hesitate to add more callbacks. Use
client/agents/agents:wait_for_requests instead of
client/civclient:wait_till_request_got_processed.